# Set Up Your Account
Source: https://docs.mira.day/en/docs/get-started/create-your-account
> Create a MiraDay account, no invitation needed. Sign up with Google, Microsoft, or your work email, verify your email, and start sourcing.
Creating a MiraDay account is open to everyone, no invitation needed. Here's how to sign up and get started.
***
Step 1: Sign up [#step-1-sign-up]
Go to [mira.day](https://www.mira.day) and click **Start for free**. You'll see the **Sign in or Sign up** screen, with three ways to create your account:
* **Continue with your Google account.**
* **Continue with your Microsoft account.**
* **Sign up with your work email.** Enter your email and click **Continue**, then create a password with at least 8 characters, including letters, numbers, and a symbol, and click **Continue** again. If your company uses single sign-on, MiraDay detects it and sends you to your company's sign-in instead.
Step 2: Verify your email [#step-2-verify-your-email]
If you signed up with email, MiraDay sends a verification code to that address. Enter the code to confirm your email. A verification code only confirms your address, it isn't how you sign in later.
Step 3: Tell us a little about yourself [#step-3-tell-us-a-little-about-yourself]
Before you start, MiraDay asks a few quick questions on one screen:
* **What's your name?** Your first and last name. If you signed up with Google or Microsoft, these are filled in from your account.
* **What best describes your role?**
* **How did you hear about MiraDay?**
* **Add your LinkedIn profile (optional).**
Click **Get started** to finish. If your account is managed by your organization through single sign-on, you skip this step.
Step 4: Start sourcing [#step-4-start-sourcing]
You'll land on the New channel screen, ready to create your first sourcing task.
For later sign-ins, go to [mira.day](https://www.mira.day) and click **Log In**.
Sign in the same way you signed up: Google, Microsoft, or your email and password. You'll land on the New channel screen.
Choose your language [#choose-your-language]
MiraDay is available in multiple languages. Pick your language from the selector on the sign-in screen, or change it anytime from the user menu at the bottom of the sidebar.
What's next [#whats-next]
* [Run Your First Search](/en/docs/get-started/quick-start): Open a channel, describe a role, and review your first candidate pool.
---
# Export & Share
Source: https://docs.mira.day/en/docs/get-started/export-and-share
> Get the value out of your candidate pool. Reveal email and phone, export an Excel file, and share candidates with your ATS, a spreadsheet, or a hiring manager.
You've reviewed your pool. Now get the value out: reveal the contacts you need, export the candidates, and share them with your team.
Reveal contact details [#reveal-contact-details]
Each candidate links to LinkedIn and the other platforms the search drew on, such as GitHub or Hugging Face, where that data exists. Email and phone are hidden by default. Click **Get email** or **Get phone** on a candidate to reveal them on demand; once revealed, they stay visible for that candidate. To do several at once, select candidates with their checkboxes and click **Get emails** or **Get phones**, up to 50 at a time. If you run out of credits partway, Mira stops, keeps what it already retrieved, and tells you.
Revealing contacts uses credits, and phone costs more than email. You can see how many a channel used in **Channel Details**, under **Credits Consumed**.
Export a data file [#export-a-data-file]
Every pool comes with a downloadable data file, an Excel (.xlsx) spreadsheet. There's no CSV option.
* **Export the full pool**: click the download icon on the data file card (it shows **Download** on hover), or open the file and click **Download profiles** at the top of the panel.
* **Export only some candidates**: select the candidates you want with their checkboxes, then click **Bulk download**. Switch to **List view** first if you want to scan the pool while you pick.
Before the file downloads, the **Download search result** dialog asks whether to **Include already enriched contact information**. It's checked by default. Uncheck it and the Email and Phone columns stay in the file but are left empty. Click **Confirm** to download.
What the file includes [#what-the-file-includes]
The file always has these columns: **Name**, **Location**, **Current Company**, **Current Position**, **Years of Exp**, **Recent Work History**, **Education**, **Qualifications** (the fit label), **Email**, and **Phone**. Link columns, **LinkedIn**, **GitHub**, **Twitter**, **Hugging Face**, and **Other Link**, appear only when at least one candidate in the file has that link.
In the Email and Phone columns, a candidate's multiple emails or phone numbers share one cell. Contacts you haven't revealed show **Not retrieved**.
Share with your team [#share-with-your-team]
Once you have the file, share it however your team works:
* **Import to your ATS**: many ATS platforms accept Excel imports.
* **Send to a hiring manager or client**: share the file directly, or send the channel link. Opening the share panel creates the link, ready to copy. The recipient signs in to MiraDay to see the conversation, the candidate cards, and the task summary; before signing in, they see only your name and the channel name.
* **Track in a spreadsheet**: drop the file into your own sheet.
* **Save to Contacts**: click **Add to contacts** on the candidates you want to keep, or select several and use the same button in the bottom bar. They're added to [Contacts](/en/docs/tools-and-teamwork/contacts) for enrichment and stay in your candidate pool.
What's next [#whats-next]
* [Refining Your Search](/en/docs/sourcing/refine-and-export): refine the pool by talking to Mira or editing the Ideal Candidate Profile.
* [Contacts](/en/docs/tools-and-teamwork/contacts): keep the people you save in one place.
---
# Run Your First Search
Source: https://docs.mira.day/en/docs/get-started/quick-start
> Run your first sourcing task with MiraDay. Describe a role, confirm the details and the Ideal Candidate Profile, and review a ranked candidate pool with profiles and contact info.
This guide walks you through your first sourcing task with MiraDay, from describing a role to reviewing a pool of qualified candidates.
Describe the role [#describe-the-role]
After signing in, you'll see the New channel screen with a text input and suggested scenarios.
You have two options:
* **Type your own description**: Paste or upload a job description, or describe the role in your own words. Be as specific as you like. For example:
> I'm looking for a senior full-stack engineer with React and Node.js experience, preferably from a high-growth startup. Location: Berlin or remote.
* **Pick a suggested scenario**: Click one of the example prompts below the input to get started quickly.
Then click the send button at the right of the input.
A formal job description is not necessary. You can describe the role the way you'd brief a colleague; Mira understands natural language, including implicit criteria like "high-growth company background" or "has built a team from scratch."
Confirm the details (only if asked) [#confirm-the-details-only-if-asked]
Mira analyzes your input and decides whether it has enough to run the search. If your description is already specific enough, Mira moves ahead without one. If key details are missing or ambiguous, Mira shows a short clarification form. Its questions are tailored to your task and focus on the choices that shape the search, such as:
* **Key responsibilities to focus on**: which parts of the role matter most for finding the right people.
* **Scope or location coverage**: the regions or work locations the search should cover.
* **Company-type preferences**: startups, high-growth companies, or a specific kind of organization.
The exact questions vary by task. Mira pre-selects reasonable defaults, so you can review, adjust if needed, and click **Submit**, or click **Skip (use defaults)** to go with Mira's recommendations. Mira uses your answers to draft the [Ideal Candidate Profile](/en/docs/meet-mira/how-mira-works) you'll confirm next.
The form only shows up when it would improve results. A thorough description up front means fewer interruptions later.
Confirm the Ideal Candidate Profile [#confirm-the-ideal-candidate-profile]
Mira turns your brief into an **Ideal Candidate Profile**: must-haves, nice-to-haves, and things to avoid. Check it, click **Edit** to change any row, then click **Search**. Mira starts sourcing only after you confirm.
Review your candidates [#review-your-candidates]
Mira searches across its talent sources and returns your candidate pool.
You'll see:
* **Candidate cards** on the right panel, each showing:
* Name, with links to their profiles on the platforms the search drew on, such as LinkedIn, GitHub, or Hugging Face.
* Current title, company, and time in role.
* Location and highest level of education.
* A **Qualifications** section that assesses how well the candidate fits your must-haves, with an overall fit label: **Good Match** (green), **Partial Match** (yellow), or **Low Match** (grey).
* Email and phone, hidden by default. Click **Get email** or **Get phone** to reveal them on demand, which consumes credits.
* **A downloadable data file** with every candidate's details. See [Export & Share](/en/docs/get-started/export-and-share) to reveal contacts in bulk, download the file, and share it.
Click a card to open the candidate's full profile, or click the LinkedIn icon next to their name to open LinkedIn.
Refine or continue [#refine-or-continue]
After reviewing the pool, keep talking to Mira in the same channel: ask for more candidates or change the criteria ("Include candidates in Amsterdam too"). You can also expand the **Ideal Candidate Profile** bar at the top of the results panel, edit it, and click **Search**. For a different role, click **New channel** in the sidebar. See [Refining Your Search](/en/docs/sourcing/refine-and-export) for more.
Each channel is saved in the sidebar, so you can come back to any search anytime. A marker next to each channel shows where it stands: a pulsing dot while it's running, a hand icon when it's waiting for your confirmation, and a blue dot when it has finished and you haven't opened it yet.
What's next [#whats-next]
* [Review Your Candidates](/en/docs/get-started/review-your-candidates): How to read the pool Mira returns, the fit label, and the Qualifications.
* [How MiraDay Searches](/en/docs/sourcing/agentic-search): How to describe a role and how Mira searches across multiple data sources.
* [How MiraDay Matches Candidates](/en/docs/match-quality-and-trust/understanding-ai-matching): How Mira Reasoning Embedding interprets your requirements.
---
# Review Your Candidates
Source: https://docs.mira.day/en/docs/get-started/review-your-candidates
> Read the candidate pool Mira delivers: where results appear, how to read the fit label and Qualifications, and what to do next.
After a search, Mira delivers a candidate pool: a ranked set of candidates who match your requirements. Here's how to read it and decide what to do next.
Where your candidates appear [#where-your-candidates-appear]
Your search results show up in two places:
* **In the conversation**: Mira summarizes what it found and gives you a downloadable data file.
* **In the right panel**: your candidates, ordered by how well they fit. A toggle at the top of the panel switches between two layouts:
* **Card view** (the default): candidate cards you can scroll through.
* **List view**: a compact table with columns for Candidate, Qualifications, Link, Email, Phone, Location, Job Title, Company, School, Degree, and Major, handy for scanning many candidates at once.
Read your first pool [#read-your-first-pool]
Each card carries a color-coded fit label, **Good Match**, **Partial Match**, or **Low Match**, next to its **Qualifications** heading, and candidates are ordered by fit, strongest at the top. The Qualifications section checks the candidate against your must-haves one by one, with the evidence behind each judgment.
The label and Qualifications are a fast read, not the whole story. The same title means different things at different companies, so check company context and career trajectory before you decide. [Inside a Candidate Profile](/en/docs/sourcing/candidate-profile) covers every card field and how to evaluate a candidate in depth.
The data file [#the-data-file]
Every pool comes with a downloadable data file. Use it to import candidates into your ATS, share them with a hiring manager, or track them in a spreadsheet. [Export & Share](/en/docs/get-started/export-and-share) covers what the file includes and how to export a custom subset.
Go deeper [#go-deeper]
* [Inside a Candidate Profile](/en/docs/sourcing/candidate-profile): what each field on a card means and how to evaluate it.
* [How MiraDay Matches Candidates](/en/docs/match-quality-and-trust/understanding-ai-matching): how the Qualifications and fit label are reasoned.
* [Refining Your Search](/en/docs/sourcing/refine-and-export): refine the pool by talking to Mira or editing the Ideal Candidate Profile.
---
# Credits
Source: https://docs.mira.day/en/docs/account/credits
> What uses credits in MiraDay, where to see how many a channel used, and how usage limits work.
MiraDay uses credits to meter the work you run in MiraDay, like running a search or unlocking a contact. This page covers what uses credits and where to see your usage.
***
What uses credits [#what-uses-credits]
* **AI model usage**: running AI-powered searches, generating candidate summaries, and other AI work.
* **Talent search**: executing a candidate search across multiple public talent data sources.
* **Contact retrieval**: unlocking a candidate's email or phone (each is a separate action).
* **Contact enrichment**: filling in missing profile details for a candidate or contact.
* **Sandbox operations**: running tasks in MiraDay's AI agent sandbox.
* **Audio transcription**: transcribing an audio file you upload, metered by the length of the audio.
Seeing how much a channel used [#seeing-how-much-a-channel-used]
Open a channel's details to see the **Credits Consumed** for that channel. When you unlock a contact's email or phone, Mira confirms that the action used credits. For your overall balance, click the credits count at the top right of the home page to open **Credits Balance**, with your remaining credits, reset date, and consumption history. The **Usage & Billing** section of **Settings** shows your current plan, credits used this period and when they reset, and your full usage details.
Plans and pricing [#plans-and-pricing]
Each plan comes with a monthly credit allowance. If you reach your limit, Mira lets you know.
| Plan | Price | Credits |
| -------------- | ------------------------------------- | -------------------------------------- |
| **Free** | $0 | 4,000 credits / month |
| **Individual** | $199 / month | 25,000 credits / month |
| **Team** | $399 / seat / month (minimum 2 seats) | 45,000 credits / seat / month |
| **Enterprise** | Custom | Custom credits and per-user allocation |
Free covers the full sourcing agent, including importing contacts and file management. Individual adds a higher credit allowance and email support. Team adds seat and member management. Team usage analytics, shared Channels, Contacts, and Files, and shared Skills and Connectors are coming soon. Enterprise adds custom credits with per-user allocation; SSO/SAML sign-in, SOC 2 reports, dedicated onboarding, and ATS integration are coming soon. Open the Plans page in MiraDay for the full feature comparison.
Credits and plan contents are indicative and may change over time; your actual entitlements prevail.
What's next [#whats-next]
* [Sign-in & Account](/en/docs/account/sign-in-and-account): manage how you sign in and your account.
* [Languages](/en/docs/account/languages): the languages MiraDay supports and how to set yours.
---
# Languages
Source: https://docs.mira.day/en/docs/account/languages
> The four languages MiraDay supports, how to set your preferred one, and what does and doesn't get translated.
MiraDay's interface and documentation are available in four languages: English, Japanese, Simplified Chinese, and Traditional Chinese. On your first visit, MiraDay matches your browser language if it's one of these four; otherwise it shows English.
***
Set your language [#set-your-language]
**In the app:** click your **profile** at the bottom-left of the sidebar, then choose **Language**. The change takes effect immediately.
**On mira.day:** use the **language selector** at the bottom of the homepage, or at the top of the sign-in page.
Supported languages [#supported-languages]
| Language | Code |
| -------------------------- | ----- |
| English | en |
| 日本語 (Japanese) | ja |
| 简体中文 (Simplified Chinese) | zh |
| 繁體中文 (Traditional Chinese) | zh-TW |
What gets translated, and what doesn't [#what-gets-translated-and-what-doesnt]
* **Interface:** all menus, buttons, navigation, and system messages appear in your selected language.
* **Mira's replies:** Mira replies in the language you write in, including languages the interface doesn't offer, and you can switch mid-conversation just by writing in another one.
* **Candidate data:** always shown in its original language, never machine-translated, so names, titles, and companies read exactly as they do at the source, without errors introduced by translation.
* **Always in English:** MiraDay, Mira, MRE, and a few other standard product and technical names.
What's next [#whats-next]
* [Sign-in & Account](/en/docs/account/sign-in-and-account): manage how you sign in and your account.
* [Credits](/en/docs/account/credits): what uses credits and where to see a channel's usage.
---
# Sign-in & Account
Source: https://docs.mira.day/en/docs/account/sign-in-and-account
> The ways to sign in to MiraDay, resetting your password, and managing your account.
This page covers how to sign in to MiraDay and manage your account. If you're setting up for the first time, start with [Set Up Your Account](/en/docs/get-started/create-your-account).
***
Ways to sign in [#ways-to-sign-in]
MiraDay uses one combined **Sign in or Sign up** page. Choose **Continue with Google** or **Continue with Microsoft** to use an existing account, or enter your email and Mira takes you to the right next step:
* **Email and password.** New to MiraDay? You'll set a password (at least 8 characters, with letters, numbers, and a symbol). Returning? Enter your password. The one-time email code you may see only verifies a new or unverified email address; it isn't a way to sign in.
* **Enterprise single sign-on (SSO).** If your organization uses SSO, entering your work email shows **Enterprise account detected**. Click the **Continue with** button to use your company's sign-in. No separate password needed.
Resetting your password [#resetting-your-password]
Forgot your password? On the password step, click **Forgot password**. Mira emails you a reset link, and you set a new one. This is also how you change your password; there's no separate change-password screen.
Managing your account [#managing-your-account]
Open the user menu from your profile at the bottom-left of the sidebar. From there you can:
* Switch your **Language**. See [Languages](/en/docs/account/languages) for the full list.
* Open **Settings**, a popup with sections for Account, General, Usage & Billing, Skills, Connectors, and Browser Extension (desktop only). Account is where you update your profile and email, upload an avatar, and delete your account. General holds language and theme (Light, Dark, or System). Team admins manage their team by clicking **Manage team** in **Usage & Billing**.
* **Sign Out**.
Your display name shows on your avatar; you can change it anytime in the **Account** section of **Settings** (**Nickname**).
Deleting your account [#deleting-your-account]
Open **Settings**, go to **Account**, and click **Delete account**. You confirm by typing your email address and then **Permanently delete**; this removes your account and all task data and can't be undone. If you have an Individual subscription, deleting your account cancels it first. If you're the admin of a Team or Enterprise organization, you can't delete your account until the organization is handled; email [support@mira.day](mailto:support@mira.day) for help. Your data is handled as described in the full privacy policy at [mira.day/privacy](https://mira.day/privacy); see also [Privacy & Compliance](/en/docs/match-quality-and-trust/privacy-policy).
What's next [#whats-next]
* [Languages](/en/docs/account/languages): the languages MiraDay supports and how to set yours.
* [Credits](/en/docs/account/credits): what uses credits and where to see a channel's usage.
---
# Privacy & Compliance
Source: https://docs.mira.day/en/docs/match-quality-and-trust/privacy-policy
> How MiraDay collects, uses, and protects candidate and user data, and the privacy regulations it meets.
This page summarizes how MiraDay handles data and the privacy regulations it's built to meet. It's a summary for quick reference. The full, authoritative policy lives at [mira.day/privacy](https://www.mira.day/privacy).
***
What we collect [#what-we-collect]
| Data type | Purpose |
| --------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Account information** | Name, email, company, role, to create and manage your account. |
| **Usage data** | Channels created, searches performed, features used, to operate and improve the product. |
| **Candidate data you upload** | Résumés and candidate information you import, to power your own searches. |
| **Candidate data from public sources and data providers** | Public résumés and candidate data, plus professional contact data licensed from data providers, that Mira searches, filters, and processes under applicable lawful bases to provide recruiting services. |
| **Mira Browser Extension data** | While the extension is active: the content of pages you visit (including pages you're signed in to), screenshots, URLs and navigation, form inputs (password fields are masked), operation logs, and session cookies, to carry out the tasks you ask for. |
How we use your data [#how-we-use-your-data]
* To provide and improve the MiraDay platform.
* To refine matching quality over time based on your usage.
* To communicate about your account and product updates.
* To meet legal obligations.
What we don't do [#what-we-dont-do]
* We don't sell your personal data.
* We don't share your private candidate data with other users.
* We don't use your candidate or user data to train AI models, and our third-party AI providers are contractually prohibited from retaining your data for training.
* We don't make hiring decisions automatically. Results are for your review, with a human in the loop.
Regulations and your rights [#regulations-and-your-rights]
The full policy at [mira.day/privacy](https://www.mira.day/privacy) is written to meet the GDPR and UK GDPR (with the UK Data Protection Act 2018), the CCPA/CPRA and other US state privacy laws (Virginia, Colorado, Connecticut, and Texas), and the EU AI Act. It also lists the retention period for each category of data.
Candidates, including people who are not MiraDay users, can request access to or deletion of their data through the contact listed in the full policy. Verified deletion requests are answered within 30 days, and deleted candidates are added to a permanent opt-out list so their data isn't processed again.
What's next [#whats-next]
* [What MiraDay Can & Can't Do](/en/docs/match-quality-and-trust/what-mira-can-and-cant-do): the human-in-the-loop limits behind these commitments.
* [Where Candidates Come From](/en/docs/match-quality-and-trust/where-candidates-come-from): the public data Mira processes.
---
# How MiraDay Matches Candidates
Source: https://docs.mira.day/en/docs/match-quality-and-trust/understanding-ai-matching
> Why Mira reasons about your requirements instead of matching keywords, how it scores each candidate against your criteria with evidence, and where the technology comes from.
On most platforms, searching for candidates means keyword matching: you type "Python developer Berlin" and get everyone with those words on their profile. It's fast, but it misses a lot, because the best candidate rarely describes themselves in your exact words.
Mira works differently. Mira Reasoning Embedding (MRE), its reasoning-based retrieval model, reads your requirements, reasons about what you actually need, and surfaces the candidates whose real career history fits. Mira then scores each one against your criteria. The examples below show what that looks like in practice.
***
What reasoning catches that keywords miss [#what-reasoning-catches-that-keywords-miss]
* **"5 years of backend experience."** Keyword search looks for "5 years" in the text. Mira works out years of experience from the dates in the candidate's work history, so the requirement is judged on the actual timeline, not on whether the profile happens to say "5 years."
* **"Someone who's built a product from 0 to 1."** Keyword search looks for the literal phrase. Mira recognizes the pattern: joining a company early and staying through its growth.
* **"High-growth company background."** Keyword search can't do this at all. Mira reads through the candidate's work history and reasons about the company's stage and trajectory.
* **"Public company plus startup experience."** Keyword search treats these as two unrelated filters. Mira reads through the whole career arc and finds people who did both at different points, which signals a specific kind of versatility.
Describe roles the way you'd tell a colleague [#describe-roles-the-way-youd-tell-a-colleague]
Because Mira reasons, you can describe a role in plain language:
> *"I need someone senior, technical but customer-facing, who's done enterprise sales at a startup that actually had product-market fit, not one that just raised a lot of money."*
Mira unpacks that into separate criteria and applies each one. No keyword combinations, filters, or search syntax.
How Mira scores each candidate [#how-mira-scores-each-candidate]
Mira doesn't hand back a black-box score. For each candidate, Mira goes through the must-have criteria in the Ideal Candidate Profile you confirmed, one at a time, and marks each one with a check (met), a dash (uncertain), or an X (not met), with a line of evidence pulled from that candidate's own history. If the profile is silent on something profiles often leave out, like team size or project scale, Mira marks it uncertain rather than guessing. If it's missing something a matching profile would normally list, like a required skill, title, or location, Mira counts that criterion as not met.
From those per-criterion results, each candidate gets an overall label, **Good Match**, **Partial Match**, or **Low Match**, so you can scan the pool fast and still see the reasoning behind every call. A Good Match meets nearly all your criteria and fails none; a Low Match fails most of them; everything in between is a Partial Match. There are no numeric scores to decode.
Where the technology comes from [#where-the-technology-comes-from]
MRE comes out of Mira's own research into reasoning-based retrieval. As of July 2026, it ranks first overall on BRIGHT, a public benchmark for reasoning retrieval, and first among embedding-based retrievers. Mira also built PJBenchmark, a recruiting-specific evaluation set built on real hiring data. Four papers describe the approach:
* [Time-Scaling Is What Agents Need Now](https://arxiv.org/abs/2601.02714)
* [Refine Thought](https://arxiv.org/abs/2511.13726)
* [PJB](https://arxiv.org/abs/2603.17386)
* [CRE-T1 Preview Technical Report](https://arxiv.org/abs/2603.17387)
What it doesn't do [#what-it-doesnt-do]
Mira won't invent a candidate who isn't there, and it doesn't make the hiring decision. It surfaces the best available matches with the evidence, and the final call is always yours. For the full picture of what to rely on Mira for and what stays with you, see [What MiraDay Can & Can't Do](/en/docs/match-quality-and-trust/what-mira-can-and-cant-do).
Getting better results [#getting-better-results]
The quality of matching depends on the quality of your input:
* Be specific, and you get sharper matches.
* Include implicit criteria, like company stage or career pattern, to give Mira more to reason about.
* Refine after you see results; each round gets sharper.
See [How MiraDay Searches](/en/docs/sourcing/agentic-search) for practical tips.
---
# What MiraDay Can & Can't Do
Source: https://docs.mira.day/en/docs/match-quality-and-trust/what-mira-can-and-cant-do
> The limits that make MiraDay trustworthy, what it reliably does, what it deliberately won't, and why the hiring decision always stays with you.
Mira handles much of the sourcing work, but the hiring judgment stays with you. Here's what you can rely on Mira for, and what it leaves to you.
***
What you can rely on Mira for [#what-you-can-rely-on-mira-for]
* **Reasoning over your criteria.** Mira reads what you're looking for and evaluates each candidate against it with evidence, not keyword matches.
* **Working across broad public data.** It draws on multiple public talent data sources and its own index, so a fit can surface from evidence spread across them.
* **Being honest about uncertainty.** When a profile is silent on something profiles often leave out, Mira marks it uncertain instead of assuming it's met.
What Mira won't do [#what-mira-wont-do]
* **It won't invent candidates or facts.** If the right person isn't in the data, Mira tells you that rather than making one up, and it never fabricates an employer, skill, or project a profile doesn't show.
* **It won't decide who to hire.** Every result is a suggestion for your review. Judging, deciding, and reaching out stay with you.
* **It won't filter or rank on protected characteristics.** Mira is designed not to screen, sort, or score people on things like age, gender, race, religion, nationality, or disability.
* **It won't infer sensitive traits.** No emotion or sentiment analysis, and no scoring on characteristics unrelated to the role.
* **It won't grab contact details on its own.** Emails and phone numbers are fetched only when you ask for them.
Accuracy depends on the data [#accuracy-depends-on-the-data]
Mira's matches are only as good as the public information available. A profile that spells out someone's work, skills, and history can be assessed precisely; a sparse one can't, and Mira will be less certain about it. If a search returns fewer people than you asked for, loosen a requirement or review the whole pool rather than just the top result.
You're in control [#youre-in-control]
You steer the process from both ends. Before a search, Mira proposes an Ideal Candidate Profile and you confirm it, so it's evaluating against criteria you approved. It stays at the top of the results, where you can edit it and search again. After the search, you review the evidence, make the call, and decide who to contact. Mira surfaces candidates and explains its reasoning; the decision is yours.
Your compliance obligations [#your-compliance-obligations]
Some jurisdictions place additional obligations on employers who use automated tools in hiring. For example, New York City's Local Law 144 requires an independent bias audit and candidate notice when an automated employment decision tool is used to screen candidates. Whether rules like these apply depends on how you use MiraDay and where you hire, so assess them for your own workflow.
What's next [#whats-next]
* [How MiraDay Matches Candidates](/en/docs/match-quality-and-trust/understanding-ai-matching): how the per-criterion, evidence-based assessment works.
* [Where Candidates Come From](/en/docs/match-quality-and-trust/where-candidates-come-from): the public data behind the matches.
* [Privacy & Compliance](/en/docs/match-quality-and-trust/privacy-policy): how Mira handles candidate and user data.
---
# Where Candidates Come From
Source: https://docs.mira.day/en/docs/match-quality-and-trust/where-candidates-come-from
> The public talent data and continuously updated index Mira reasons across, how broad the coverage is, and how to read a profile's freshness.
Mira reasons over the best available public data about people and companies. It doesn't rely on a single source; it draws on several kinds of public professional data, then evaluates what it finds against your criteria. Here's where that data comes from.
***
What Mira draws on [#what-mira-draws-on]
| Source | What it covers |
| -------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Public talent data** | Aggregated from public platforms such as LinkedIn, X, GitHub, and Hugging Face, plus academic papers and company information, and from data providers such as Exa, Apollo, and RocketReach. |
| **MiraDay's own index** | A continuously updated index of public professional profiles and job postings. |
| **Mira Browser Extension** | For sites behind a sign-in or not in the index, [the extension](/en/docs/tools-and-teamwork/mira-browser-extension) navigates them in your own browser, on demand. |
Mira reasons across all of these together, so a candidate can surface from evidence spread over more than one source, not just whoever happens to match a single database.
How broad the coverage is [#how-broad-the-coverage-is]
Coverage is global, across industries, regions, and seniority levels, from early-career to executive. MiraDay's index spans 4.5 billion public records. For any one search, what matters is whether the people who fit your criteria are in the data, not the raw size.
How current the data is [#how-current-the-data-is]
Public data ages: people change jobs, and profiles aren't always updated the day they do. When you open a search step in the conversation, each person in that step's results shows a **freshness badge** marking when their data was last updated, for example "Updated · 3w ago", with the exact date on hover. When an update date isn't available, no badge is shown rather than a guess, and profiles updated within the last day don't show one either.
Use it as a signal: a very fresh profile is more likely to reflect someone's current role, while an older one is worth a second look or a quick confirmation before you reach out.
What this means for your search [#what-this-means-for-your-search]
Because Mira reasons over evidence rather than pulling from one list, two things follow. If a search returns fewer people than you asked for, it's because fewer people in the available data met your must-have requirements, not because of a hidden cap. Loosen a requirement and you'll usually see more. And accuracy tracks how complete the public data is: the more a profile actually says, the more precisely Mira can assess it. For how that assessment works, see [How MiraDay Matches Candidates](/en/docs/match-quality-and-trust/understanding-ai-matching).
What's next [#whats-next]
* [How MiraDay Matches Candidates](/en/docs/match-quality-and-trust/understanding-ai-matching): how Mira reasons over this data to score each candidate.
* [Inside a Candidate Profile](/en/docs/sourcing/candidate-profile): what each field on a candidate means.
* [Mira Browser Extension](/en/docs/tools-and-teamwork/mira-browser-extension): reach sites and sign-ins that aren't in the index.
---
# Changelog
Source: https://docs.mira.day/en/docs/help/changelog
> Recent updates and new features in MiraDay.
September 2026 [#september-2026]
Edit your candidate profile right in the results panel and search again in place [#edit-your-candidate-profile-right-in-the-results-panel-and-search-again-in-place]
Previously, adjusting your search meant describing it again in the conversation, and criteria you had deleted from the profile card sometimes found their way back into the next round. Now a collapsible **Ideal Candidate Profile** bar sits at the top of the results panel, showing how many must-have, nice-to-have, and avoid criteria it holds. Expand it to add, edit, or remove criteria one by one, then click **Search** to rerun the search in place with the new profile. If the previous round is still running, you're asked whether to wait for it to finish or start the new search right away. The profile you confirm is used exactly as it is, so deleted criteria stay deleted. Market insights Mira surfaces while sourcing are now labeled **Insight** throughout.
Delivered lists are more complete and cleaner [#delivered-lists-are-more-complete-and-cleaner]
Previously, candidates Mira found through the browser on recruiting platforms such as Liepin often fell away at delivery: they were missing from the list, or their cards were incomplete with no link back to the original platform. These candidates are now delivered in full like those from any other source, and their cards link straight back to the original page. When you ask to search only a specific platform, Mira keeps the search there; if it's temporarily unavailable, Mira asks before switching to another source instead of switching quietly. Mira also now holds back people whose profiles give no indication of what they do. The summary after delivery also no longer repeats what each card already shows, and instead describes the list as a whole, such as which cities and companies the candidates are concentrated in.
See every email and phone number, and choose whether to include them in downloads [#see-every-email-and-phone-number-and-choose-whether-to-include-them-in-downloads]
Previously, a candidate with several emails or phone numbers showed only one, and a second one disappeared on refresh after it was retrieved. Now every email and phone number retrieved is kept: the card shows the first with a count of how many more there are, and the details view and table show them all. Before you download a candidate list, there's a new **Include already enriched contact information** option, checked by default, with multiple values placed in the same cell. Uncheck it and the email and phone columns stay in the file but are left empty. This also fixes the candidate table not scrolling sideways on phones.
Interface languages cut from 13 to four: Simplified Chinese, Traditional Chinese, English, and Japanese [#interface-languages-cut-from-13-to-four-simplified-chinese-traditional-chinese-english-and-japanese]
The interface used to be switchable between 13 languages. It now offers Simplified Chinese, Traditional Chinese, English, and Japanese. Accounts that had picked another language will see the interface in English, and new visitors are matched to their browser language. This only affects menu and button text: Mira still replies in whatever language you write in, so if you ask in Korean, it answers in Korean.
Share links now require sign-in to view [#share-links-now-require-sign-in-to-view]
Previously, anyone who opened a share link could see everything in the channel. Now a visitor who isn't signed in sees only the sharer's name and avatar and the channel name, and after signing in they're taken straight back to the link to view the full content. The share panel is also down to one step: opening it generates the link, ready to copy, with no separate **Create** click.
An outdated browser extension now prompts you to update [#an-outdated-browser-extension-now-prompts-you-to-update]
When the extension is below the required version, an update card appears with the exact steps to update: first check for updates manually on your browser's extensions page; if that doesn't work, remove the extension and click **Open store** on the card to reinstall it. With more recent versions of the extension, you can instead click **Update now** to update in one click without leaving the page.
Choose how many candidates to deliver, and pick more than one experience range [#choose-how-many-candidates-to-deliver-and-pick-more-than-one-experience-range]
You can now enter the number of candidates you want in the clarification form, or simply state it in your brief (up to 100 per search), and Mira delivers that many. The years-of-experience question also accepts several ranges at once, such as "1–3 years" and "8–10 years."
Wider search first, then delivery ranked by relevance [#wider-search-first-then-delivery-ranked-by-relevance]
Previously each source was read only to its first page, so the people at the top weren't necessarily the best fit for the brief. Mira now pulls several pages from each source, removes duplicates, ranks the pool by relevance to the brief, and only then enriches and evaluates the top of the list. Evaluation now runs in batches and is noticeably faster. A brief that covers several roles is sourced separately per role instead of being merged into one list.
Fixed: enterprise single sign-on bouncing back to the sign-in page [#fixed-enterprise-single-sign-on-bouncing-back-to-the-sign-in-page]
Previously, some users signing in through enterprise single sign-on (SSO) were sent back to the sign-in page even though sign-in had succeeded, with no message and no luck on retry. This is fixed. If an account really has been disabled, the sign-in page now says so clearly instead of silently returning you.
August 2026 [#august-2026]
Contact buttons that wouldn't respond now work [#contact-buttons-that-wouldnt-respond-now-work]
Previously, some candidate cards kept showing **No Email** or **No Phone** and wouldn't respond to a click even though the contact details could be retrieved, while others failed every time. A candidate's reveal state and their profile were kept in two separate places that often didn't line up. The two are now merged into one, and **Get Email** and **Get Phone** work as expected for those candidates.
More reliable delivery of your candidate list [#more-reliable-delivery-of-your-candidate-list]
A search used to run into several problems at delivery: someone on the delivered list couldn't be found anywhere in your candidate pool; a list that should have been cards rendered as a wall of raw code; a search that found nobody returned an error instead of an empty state; a preview flashed "unsupported file format" while it was still loading; and one class of delivery failed outright, leaving only text in the panel. These are all fixed, and everyone on a delivered list can now be found in your pool and followed up on.
Hard requirements in your brief are now actually enforced [#hard-requirements-in-your-brief-are-now-actually-enforced]
Previously, writing things like "only people currently employed," "must have a bachelor's degree or above," or "must have worked at this kind of company" often got dropped somewhere along the way, and people who didn't meet them still showed up. These requirements now carry through all three stages (search, filtering, and evaluation), so people who don't meet them no longer reach the delivered list.
Tell Mira to find another way when it hits a login wall [#tell-mira-to-find-another-way-when-it-hits-a-login-wall]
When Mira used a browser and reached a site that required a login or a CAPTCHA, the card offered only **Continue**, so you either handled it in the browser yourself or sent another message to work around it. There's now a **Cancel** button: click it and Mira picks a different source or a different approach on its own instead of waiting for you. This also fixes the card previously showing a false red "action failed" message.
See how far along a task is from the sidebar [#see-how-far-along-a-task-is-from-the-sidebar]
Previously you had to open a task to know whether it had finished. The task list in the sidebar now shows a spinner for a task that's running and a dot for one that finished while you weren't looking, and it stays in sync across your devices and browser tabs.
Images and PDFs you send now reach Mira properly [#images-and-pdfs-you-send-now-reach-mira-properly]
Previously an image or PDF sent in a conversation sometimes never reached Mira at all, and sometimes returned an error. Attachments now come through. A related problem, where PDFs were lost when forwarded alongside other file types in email, is fixed as well.
A very long PDF no longer freezes the whole conversation [#a-very-long-pdf-no-longer-freezes-the-whole-conversation]
Previously, once the PDFs uploaded across a conversation added up to too many pages, every following turn in that conversation came back blank. Retrying didn't help, and you had to start over. Mira now switches to a different way of reading those files, and the conversation carries on.
More complete candidate cards, without duplicates [#more-complete-candidate-cards-without-duplicates]
Previously some candidates' work history showed only the company and title, with no dates or duration, and one person's LinkedIn profile could appear as two separate links because of regional URL differences. Dates and duration now show in full, and a profile appears only once.
Refreshed home scenario cards and interface styling [#refreshed-home-scenario-cards-and-interface-styling]
The scenario cards on the home page have a new version, each with a title, a description, and a few prompts you can use straight away. Channel, the task page, Settings, Contacts, and candidate detail pages also went through a round of styling work, along with fixes to mobile layout.
Get contact details for several candidates at once [#get-contact-details-for-several-candidates-at-once]
Previously you could only reveal a candidate's email and phone one person at a time. Now you can multi-select candidates in the list and pull email and phone for several of them in one go (up to 50 per batch). When your remaining balance runs low, you get a clear message instead of a vague "couldn't retrieve."
Send email with attachments in Mira [#send-email-with-attachments-in-mira]
Once you've connected Gmail or Outlook, you can now add attachments when composing and forwarding email. When you forward a message, its original attachments come along, and you can add your own on top.
A refreshed way to follow your Agent Team's collaboration [#a-refreshed-way-to-follow-your-agent-teams-collaboration]
When an Agent Team works, the whole collaboration now unfolds like a group chat: each teammate appears as its own participant in the conversation, showing its status, messages, and deliverables in real time. A new task workspace panel on the right gathers the team's output in one place.
Paste or drag files in to attach them [#paste-or-drag-files-in-to-attach-them]
Previously you could only add attachments by clicking the plus button and picking from a file selector, and a screenshot had to be saved to a file first. Now you can paste an image straight from your clipboard into the input box, or paste and drag files in.
July 2026 [#july-2026]
Start using MiraDay right after signing up, no invite code needed [#start-using-miraday-right-after-signing-up-no-invite-code-needed]
Previously new users had to activate with an invite code before they could start. Now activation happens as soon as you finish signing up, so you can jump straight in without an invite code.
Cleaner contact details and profiles for candidates [#cleaner-contact-details-and-profiles-for-candidates]
A candidate's email and phone now appear as clickable chips, and clicking a candidate card opens their full profile in a side panel, with clear confirmation when you copy a contact detail.
Settings in one popup you can open anywhere [#settings-in-one-popup-you-can-open-anywhere]
Previously changing a setting meant navigating to a separate settings page. Now you can open the settings popup from anywhere, organized into Account, General, Usage & Billing, Team, Skills, and Connectors, and you can upload an avatar, change your email, view usage details, and delete your account.
Richer, clickable sources and supporting evidence for candidates [#richer-clickable-sources-and-supporting-evidence-for-candidates]
A candidate's card now gathers that person's source profiles in one place (LinkedIn, GitHub, Hugging Face, Twitter, and recruiting platforms), and you can click straight through to the original page. The **More** section also shows supporting evidence from the web and news (article titles, relevant quotes, and publish dates), along with public profiles from platforms like GitHub and Hugging Face, so you can vet a candidate's background more thoroughly.
Fixed searches that delivered incomplete or empty candidate lists [#fixed-searches-that-delivered-incomplete-or-empty-candidate-lists]
Previously some searches had actually found candidates, yet the final delivered list came back empty or incomplete (especially for people found through the browser extension) while the task still showed as finished. This is now fixed, and the people that are found are delivered in full.
Switch candidate results between card view and table view [#switch-candidate-results-between-card-view-and-table-view]
Candidate search results can now switch between a card view and a table view: the table view makes it easy to compare candidates side by side, and the card view makes it easy to read one candidate's full résumé in detail.
Fixed tasks that kept spinning after they finished [#fixed-tasks-that-kept-spinning-after-they-finished]
In some cases a task had actually finished or stopped, but the interface kept showing the loading spinner. Task status now updates promptly and correctly.
File downloads no longer show garbled text on iPhone and iPad [#file-downloads-no-longer-show-garbled-text-on-iphone-and-ipad]
Fixed an issue where downloading a file in iOS Safari could garble non-Latin file names or file contents. Downloaded file names and contents now display correctly.
Contacts: keep the people you find [#contacts-keep-the-people-you-find]
Save candidates beyond a single search. Click **Add contact** on a result, or **Import** your own people from PDF, DOCX, CSV, or XLSX files, and they gather in a new **Contacts** page on the sidebar. Each contact has a full profile you can enrich, reveal email and phone, and add notes, so the people you find don't get lost when the channel ends.
One person, one card across sources [#one-person-one-card-across-sources]
When the same person turns up in more than one data source, Mira now links those results into a single candidate, instead of showing the same person several times. Your pool reads cleaner and the count reflects real, distinct people.
June 2026 [#june-2026]
Upload audio files [#upload-audio-files]
You can now attach audio files up to 100 MB, such as meeting recordings, role-requirement calls, and interview recordings. Mira listens to them and works from what it hears, continuing with your task.
Files: every file from your channels in one place [#files-every-file-from-your-channels-in-one-place]
The résumés and JDs you upload and the data files and reports Mira generates now live together on a dedicated Files page, reachable from the sidebar, instead of being scattered across separate channel threads. Filter by source, type, or channel, search by name, and jump back to the channel any file belongs to.
Connect your Gmail and Outlook [#connect-your-gmail-and-outlook]
Open the Connectors entry in your personal settings or the channel input area to connect your own Gmail or Outlook account. Once connected, Mira can handle email correspondence directly through your mailbox while working on a channel, with no need to switch back and forth between Mira and your inbox.
Browser-extension sourcing returns the full number of candidates you asked for [#browser-extension-sourcing-returns-the-full-number-of-candidates-you-asked-for]
When sourcing candidates on hiring platforms with the browser extension, larger requests sometimes stopped early and returned fewer candidates than you asked for. This is now fixed: browser-extension sourcing reliably completes the request you set.
Special characters no longer freeze the page [#special-characters-no-longer-freeze-the-page]
Fixed an issue where certain unrecognized special characters caused the page to go blank. When you type text containing angle-bracket tags or other special characters in the input box, Mira now treats them as plain text and handles them correctly, so input is more stable.
Agentic Search: autonomous source planning across data sources [#agentic-search-autonomous-source-planning-across-data-sources]
When searching for candidates, Mira can now plan its own retrieval path across multiple data sources and run the search autonomously. Paired with an Agent Team it handles complex searches better, and its parallel processing returns results faster.
Agent Team takes on complex tasks together [#agent-team-takes-on-complex-tasks-together]
An Agent Team is made up of several teammates that can collaborate and communicate on their own. Once you give an instruction, Mira assembles the team automatically; you can also shape a specific teammate's role and how independently it works, from fully automatic to review-before-send or step-by-step confirmation. It suits long-running deep research and complex hiring tasks. You can turn it on from the button in the channel input, and each teammate's role and progress appears in the channel conversation in real time.
Skills are live [#skills-are-live]
You can now save frequently used workflows as personal skills and invoke them by typing / in the channel input, instead of describing the same requirements from scratch each time. The official skill-creator is live too. It lets you create and edit personal skills through conversation.
Candidate search follows your requirements more strictly [#candidate-search-follows-your-requirements-more-strictly]
Before a search starts, Mira now drafts an ideal-candidate profile for you to review and adjust, and only begins searching once the direction is confirmed. Hard requirements you set, such as specific experience or target companies, are strictly enforced, so candidates who fail them no longer appear in results. All retrieved candidates are also evaluated in full, rather than being cut off early by a count limit.
File previews load faster and cover more file types [#file-previews-load-faster-and-cover-more-file-types]
Previously, previewing files such as PDFs and PowerPoint decks required the browser to download and render the entire file, which became slow with many files. Thumbnails are now pre-generated images, so the file panel opens noticeably faster. Code and configuration files now have previews, HTML files render as actual pages, and the selected state when multi-selecting files is clearer.
File previews always show the latest version [#file-previews-always-show-the-latest-version]
Previously, after a file in a channel was updated, its preview could occasionally still show the older content. Opening a file now always shows the most recent version.
May 2026 [#may-2026]
Upload compressed archives as attachments [#upload-compressed-archives-as-attachments]
You can now attach compressed archives such as zip, tar, gz, and rar directly to a channel. Mira unpacks them inside its sandbox and reads the contents automatically, so there's no need to extract and upload files one by one. Supplying material in bulk is more efficient.
Candidate match levels at a glance [#candidate-match-levels-at-a-glance]
Candidate match analysis is now tiered, using color-coded labels to distinguish how well each candidate fits, so you can size up suitability at a glance.
Speak instead of typing [#speak-instead-of-typing]
You can now hold the microphone button in the channel input and just talk, and Mira transcribes your voice into text automatically. Describing the kind of candidate you're looking for is easier when you can say it out loud, and the context that comes out tends to be richer.
Scroll back through more channel history [#scroll-back-through-more-channel-history]
The channel history in the sidebar now loads incrementally as you scroll. Older channels appear automatically instead of being cut off at a fixed count, so finding something you ran earlier is much easier.
More complete candidate profiles [#more-complete-candidate-profiles]
The way candidate profile data is checked has been tightened, so the work history and education shown on the candidate page come out more complete than before.
Years-of-experience filter is more accurate [#years-of-experience-filter-is-more-accurate]
The "years of experience" filter now counts a candidate's "career span", the total years from their first real job to today, instead of summing up overlapping periods. We also fixed a low-probability case where location filtering returned no results when it should have.
Your language choice now sticks across MiraDay [#your-language-choice-now-sticks-across-miraday]
Previously, moving between MiraDay's docs, blog, and the agent app sometimes meant picking your language again. Now once you set a preferred language anywhere, every other surface in MiraDay switches to it automatically the next time you visit.
Browser extension performance upgrade [#browser-extension-performance-upgrade]
The extension connects to recruiting platforms more reliably, with far fewer disconnects and connection failures. The channel input page now clearly shows whether you're connected to the target site and which account is authorized, with a visible prompt when the browser can't connect. The extension can also auto-fill and submit search forms on recruiting platforms, so Mira can complete searches end-to-end on more sites.
Candidate info is easier to scan [#candidate-info-is-easier-to-scan]
The candidate card layout has been refreshed so match analysis, skill sets, and working experience sit in more prominent positions, so you can size up whether a candidate fits at a glance.
Fewer agent interruptions [#fewer-agent-interruptions]
We fixed several causes of tasks stopping partway through, such as network disconnects and tool-call failures, so far more tasks now run to completion.
PowerPoint files render more reliably [#powerpoint-files-render-more-reliably]
The PowerPoint renderer has been reworked, fixing the occasional layout glitches in generated slides. The PowerPoint view also loads faster on first open.
April 2026 [#april-2026]
Talent data keeps expanding [#talent-data-keeps-expanding]
MiraDay's talent dataset keeps growing, covering more candidate information from public sources and more reachable contact methods.
Mira Browser Extension is live [#mira-browser-extension-is-live]
The Mira Browser Extension is now available on the Chrome Web Store. It uses your own logged-in browser session to search recruiting platforms and gather information, useful when you need to find candidates MiraDay's built-in search doesn't cover.
Share a channel with others [#share-a-channel-with-others]
Channels completed inside MiraDay can now be shared via a generated link. The recipient doesn't need a MiraDay account. They can open the link to see the full channel conversation, the generated candidate cards, and the AI task summary, making it more direct to hand off context.
---
# FAQ
Source: https://docs.mira.day/en/docs/help/general-faq
> Common questions about MiraDay: what it is, running your first search, working as an Agent Team, saving and sharing candidates, the browser extension, Skills, email connectors, access, and data.
What is MiraDay? [#what-is-miraday]
MiraDay is a Talent Intelligence platform for professional and leadership hiring. Describe a role in plain language and Mira returns a pool of qualified candidates, each with a profile, Qualifications, and contact details. See [What is MiraDay](/en/docs/meet-mira/what-is-mira) for the full picture.
How do I run my first search? [#how-do-i-run-my-first-search]
On the New channel screen, describe the role, paste or upload a job description, or pick a suggested scenario, then send. Mira may ask a few clarifying questions first, then returns a ranked candidate pool in a few minutes, and you keep refining in the same conversation. See [Run Your First Search](/en/docs/get-started/quick-start).
Can Mira work as a team on a big search? [#can-mira-work-as-a-team-on-a-big-search]
Yes. Every candidate search runs as an [Agent Team](/en/docs/meet-mira/agent-teams). Mira hands the search to a sourcing teammate and, on bigger jobs, brings in more teammates, each taking one part of the job (sourcing, screening, or matching). They work together in one channel while you stay in control.
Where do the candidates come from? [#where-do-the-candidates-come-from]
Mira aggregates data from multiple public talent data sources, plus its own index of public professional profiles. You can also upload your own resumes, and use the [Browser Extension](/en/docs/tools-and-teamwork/mira-browser-extension) to reach sites you're logged into, such as your ATS. See [What is MiraDay](/en/docs/meet-mira/what-is-mira#where-do-the-candidates-come-from) for details.
How do I keep the candidates I find? [#how-do-i-keep-the-candidates-i-find]
Save anyone beyond a single search to [Contacts](/en/docs/tools-and-teamwork/contacts), your workspace-level list. Click **Add to contacts** on a candidate, or **Import** your own people from PDF, Word, Excel, CSV, or HTML files. Mira keeps each contact's profile enriched and up to date.
How do I share candidates with my team or hiring manager? [#how-do-i-share-candidates-with-my-team-or-hiring-manager]
Reveal the email and phone you need with **Get email** and **Get phone**, then download the pool as an Excel (.xlsx) file to hand off to your ATS, a spreadsheet, or a hiring manager. Revealing contacts uses credits. See [Export & Share](/en/docs/get-started/export-and-share).
What is the Mira Browser Extension, and how do I install it? [#what-is-the-mira-browser-extension-and-how-do-i-install-it]
The [Mira Browser Extension](/en/docs/tools-and-teamwork/mira-browser-extension) lets Mira operate your own Chrome browser in a dedicated tab, using your own logins. Install it from the Chrome Web Store and authorize it once. Use it to read the latest profiles from sites like LinkedIn so candidate data stays current, reach candidates behind a sign-in, research companies on the live web, and reach other tools you're signed into such as your ATS.
Can I save workflows I use repeatedly? [#can-i-save-workflows-i-use-repeatedly]
Yes. Save a recurring workflow (an intake format, an evaluation rubric, a report structure) as a [Skill](/en/docs/tools-and-teamwork/skills), then invoke it with `/` in the channel input instead of describing the same steps each time. Official Skills come from the Mira team; personal Skills are yours to create and share.
Can I connect my email? [#can-i-connect-my-email]
Yes. Connect your Gmail or Outlook from [Connectors](/en/docs/tools-and-teamwork/connectors), and Mira can draft, send, and summarize email as part of a channel, without switching back to your inbox.
Where are my uploaded files and reports? [#where-are-my-uploaded-files-and-reports]
[Files](/en/docs/tools-and-teamwork/files) gathers everything from your channels in one place: the resumes and JDs you upload, plus the data files and reports Mira generates, so nothing is scattered across channel threads. Open **Files** from the sidebar.
How do I get access, and what does it cost? [#how-do-i-get-access-and-what-does-it-cost]
MiraDay is open to sign up, no invitation needed. Go to [mira.day](https://www.mira.day), continue with Google or Microsoft, or sign up with your email and a password, then verify your email with the code we send. Plans are Free ($0), Individual ($199/month), Team ($399/seat/month), and Enterprise (custom pricing); see [Credits](/en/docs/account/credits) for what each plan includes.
Is my data safe? [#is-my-data-safe]
Yes. Your data is encrypted in transit and at rest, and private data is accessible only within your workspace. See the [Privacy Policy](/en/docs/match-quality-and-trust/privacy-policy) for details.
**Looking for more on searching?** The [Sourcing FAQ](/en/docs/help/sourcing-faq) covers how many candidates you get, how long a search takes, filtering by location, refining results, and downloading the data.
Ready to try MiraDay? [Sign up](https://www.mira.day) and get started.
---
# Sourcing FAQ
Source: https://docs.mira.day/en/docs/help/sourcing-faq
> Frequently asked questions about MiraDay's sourcing capabilities.
How many candidates can I get? [#how-many-candidates-can-i-get]
By default Mira returns 40 candidates. To get more or fewer, say so when describing the role, for example, "find me 50 ML engineers" or "show me 10 senior PMs", or enter the number in the clarification form, up to 100 per search. You can also ask for more candidates in the same conversation at any time.
How long does a search take? [#how-long-does-a-search-take]
Most sourcing rounds complete within a few minutes. Complex searches with very specific criteria may take longer.
Can I search for candidates in specific countries or cities? [#can-i-search-for-candidates-in-specific-countries-or-cities]
Yes. Include location in your description (e.g., "Berlin or remote", "London", "Southeast Asia"). Mira understands geographic preferences and can filter or prioritize by location.
What if the results don't match what I need? [#what-if-the-results-dont-match-what-i-need]
Expand the **Ideal Candidate Profile** bar at the top of the results panel, add, edit, or remove criteria, then click **Search** to run a new round of results with the updated profile. You can also tell Mira in the same conversation, for example: "Focus more on startup backgrounds" or "Include candidates from Amsterdam too." Mira adjusts the next search based on your feedback.
Can Mira work with candidates from my ATS? [#can-mira-work-with-candidates-from-my-ats]
You can upload resume files directly, or import an ATS export (CSV or Excel) into [Contacts](/en/docs/tools-and-teamwork/contacts). With the Mira Browser Extension, Mira can also navigate your ATS in your browser using your own login. This is browser automation you control, not a direct ATS integration.
What does the Qualifications assessment tell me? [#what-does-the-qualifications-assessment-tell-me]
The Qualifications assessment is a concise evaluation of how well the candidate fits your search requirements. It opens with a color-coded label, Good Match (green), Partial Match (yellow), or Low Match (grey), then checks the candidate against your must-haves criterion by criterion and highlights relevant experience, skills, and career patterns, not a generic bio.
Can I download the candidate data? [#can-i-download-the-candidate-data]
Yes. Every candidate pool comes with a downloadable data file containing all candidate information in a structured format. You can import this into your ATS or spreadsheets.
Does Mira remember my preferences across channels? [#does-mira-remember-my-preferences-across-channels]
Within a channel, Mira keeps the full conversation context and refines based on your feedback. Each new channel starts fresh, but all your past channels stay in the sidebar to revisit or continue anytime.
---
# Troubleshooting
Source: https://docs.mira.day/en/docs/help/troubleshooting
> Common issues and how to resolve them.
I can't sign in [#i-cant-sign-in]
* Make sure you're using the email address you registered with, and sign in with your password.
* Forgot your password? Use **Forgot password** to reset it by email.
* If you signed up with Google or Microsoft, use that same button to sign in.
* If your company uses enterprise single sign-on (SSO), enter your work email and continue with your company's sign-in. No MiraDay password is needed.
* If your account has been disabled, the sign-in page says so.
* A verification code, if you receive one, only confirms a new or unverified email address. It isn't how you sign in.
Mira isn't returning results [#mira-isnt-returning-results]
* Check that your description is specific enough. Very vague queries (e.g., "find me someone") may not produce useful results.
* Try including more context: role type, skills, experience level, location.
* If a search runs longer than a few minutes, check the channel in the sidebar: a spinner means it's still running, and a dot means it finished. Refresh the page if nothing changes.
Candidate contact information is missing [#candidate-contact-information-is-missing]
* Not all candidates have publicly available email addresses or phone numbers.
* Click **Get phone** to try retrieving a phone number (this uses credits).
* Use the LinkedIn link to reach candidates through InMail.
The browser extension says it needs an update and I can't skip it [#the-browser-extension-says-it-needs-an-update-and-i-cant-skip-it]
* Your Mira Browser Extension is older than the version Mira requires, so browsing is blocked until you update.
* If the card shows **Update now**, click it to update in one click.
* Otherwise, go to `chrome://extensions/`, turn on **Developer mode**, and click **Update**.
* If it still won't update, click **Remove** on the Mira extension, then use **Open store** on the card and click **Add to Chrome** to reinstall.
* Once you've updated, Mira picks the channel back up.
The interface is in the wrong language [#the-interface-is-in-the-wrong-language]
* Open **Language** from your profile menu (click your name or avatar in the bottom-left), or open **Settings** and go to **General**.
* On mira.day, use the language selector at the bottom of the page.
Something else isn't working [#something-else-isnt-working]
If you're experiencing an issue not covered here:
1. Try refreshing the page.
2. Try signing out and back in.
If the issue persists, reach out to the Mira team for help with a description of what happened.
---
# Agent Team
Source: https://docs.mira.day/en/docs/meet-mira/agent-teams
> For every candidate search, and for other multi-step jobs, Mira assembles an Agent Team whose teammates divide the task and work it together, cross-checking and challenging each other, while you stay in control and steer as it goes.
With an Agent Team, Mira acts as team lead and brings in teammates, each taking a different part of the task, while you stay in control. Recruiting searches are the most common use, but a team can take on any hard, multi-step job.
***
A teammate for every part [#a-teammate-for-every-part]
A hard job has several moving parts. For a search, that often means several sourcing teammates working in parallel, each on its own region or talent segment, with others researching companies or checking facts. The same holds for other multi-step work, like mapping a market or profiling several companies at once.
Every candidate search runs as an Agent Team: Mira always hands the search to at least one sourcing teammate and brings in more when the work needs them. For other tasks, Mira decides; a focused, well-scoped request usually runs with Mira alone.
The team works together in the channel [#the-team-works-together-in-the-channel]
A channel is where an Agent Team does its work. Like a real team in a group chat, teammates share what they find, build on each other's work, and challenge each other's conclusions to reach a better result. You can follow along from start to finish, check any step as it happens, and trace the evidence behind every conclusion.
In the **Task workspace** on the right, **Task plan** tracks each stage as it completes, and **Member status** shows what each teammate is doing live.
When the team finishes, Mira pulls every teammate's results together and delivers the final answer to you in the main conversation.
Who's on the team [#whos-on-the-team]
When the team starts, Mira posts who it enlisted: "Mira assembled the Agent Team, enlisting N teammates," followed by each teammate's name, role, and what they're responsible for. Mira then posts "Mira invited … to the channel" and lists each teammate's **Assignments**, the specific work for this run.
Teammates work through their part on their own. When a step involves outreach, sensitive data, or a high-risk action, Mira waits for your confirmation first.
Steer the search as it runs [#steer-the-search-as-it-runs]
You confirm the direction rather than micromanage the work. Before the team searches, Mira has you review and confirm an **Ideal Candidate Profile**: your must-haves, nice-to-haves, and things to avoid. Once the team is running, you steer by talking to Mira in the channel, and Mira passes your guidance to the right teammate. You can also edit the profile in the **Ideal Candidate Profile** bar at the top of the results panel and click **Search**. If a search is still running, Mira asks whether to **Wait** or **Start new search**, which stops the current one.
Starting a team [#starting-a-team]
There are two ways a team starts:
1. **Mira assembles one automatically** for every candidate search, and for any other task that's a better fit for a team.
2. **You turn it on** with the **Agent Team** button in the channel input toolbar. It highlights when active and is remembered per channel. Turn it on to have Mira use a team for other kinds of work too; candidate searches use a team either way.
Credits [#credits]
Like other work in MiraDay, an Agent Team run uses credits. You can see how many a channel used in **Channel Details**, under **Credits Consumed**.
FAQ [#faq]
Can teammates use the browser extension? [#can-teammates-use-the-browser-extension]
Yes. If the work needs it, authorize browser control first, a one-time approval you give in the main conversation, and then teammates can browse too. They can't request that authorization mid-run, so turn it on up front. See [Browser Extension](/en/docs/tools-and-teamwork/mira-browser-extension).
Can I run a search without an Agent Team? [#can-i-run-a-search-without-an-agent-team]
No. Every candidate search runs as an Agent Team, even when it needs only one sourcing teammate. For other tasks, leave the Agent Team button off and Mira decides; a focused request usually runs with Mira alone.
What's next [#whats-next]
* [How MiraDay Searches](/en/docs/sourcing/agentic-search): the search a team runs across many sources.
* [Browser Extension](/en/docs/tools-and-teamwork/mira-browser-extension): let Mira browse on your behalf.
---
# How MiraDay Works
Source: https://docs.mira.day/en/docs/meet-mira/how-mira-works
> You describe a role and confirm an Ideal Candidate Profile. Mira's Agent Team then searches with MRE (Mira Reasoning Embedding), a reasoning-based retrieval model, and delivers a ranked candidate pool.
You describe the role and confirm what a great candidate looks like; Mira and its teammates do the searching, reasoning, and matching. Here's the workflow.
You describe the need [#you-describe-the-need]
Paste or upload a job description, upload notes from an intake meeting, or describe the role in plain language. Mira understands implicit criteria; you don't need formal search syntax.
Mira proposes an Ideal Candidate Profile [#mira-proposes-an-ideal-candidate-profile]
When key details are missing or ambiguous, Mira shows a short clarification form to confirm the requirements with you, so the search aims at the right target. Mira then enriches the brief and proposes an **Ideal Candidate Profile**, a structured summary of the search in three lists: **Must-Have**, **Nice-to-Have**, and **Avoid**. Intake and the profile are run by a built-in [Skill](/en/docs/tools-and-teamwork/skills).
You review and confirm the profile [#you-review-and-confirm-the-profile]
The Ideal Candidate Profile is yours to edit. Adjust any of the three lists: add a **Must-Have**, move something into **Nice-to-Have**, or add a competitor to **Avoid**. Then choose **Search** to accept it as-is, or edit first and then search. Mira starts searching only after you confirm, so the search runs against the bar you set. See [how Mira reads your requirements](/en/docs/sourcing/search-filters-and-criteria) for how these criteria shape results.
Mira runs the search [#mira-runs-the-search]
Once you confirm, Mira hands the search to an [Agent Team](/en/docs/meet-mira/agent-teams): at least one sourcing teammate, more when the search is complex. The search works in a loop, powered by reasoning-based retrieval (MRE): the team searches across many sources, surfaces signals, reasons over what it finds, and decides where to go next, whether to keep digging or to change its search strategy or sources. Each candidate is then evaluated against your confirmed profile, not by keyword matching:
* "At least 5 years of experience": calculated from the career timeline.
* "From 0 to 1 experience": interpreted as early-stage company building.
* "Public company background": inferred from organizational context.
Candidate pool delivered [#candidate-pool-delivered]
You get a curated candidate pool with:
* Candidate name, current role, company, and location.
* Most recent work experience.
* A Qualifications assessment: how the candidate fits your must-haves, criterion by criterion.
* Highest level of education.
* Contact info: LinkedIn is shown by default; email and phone are revealed on demand by clicking **Get email** or **Get phone** (each reveal uses credits).
* A downloadable data file.
Iterate [#iterate]
Continue the conversation:
* Ask for more candidates.
* Adjust criteria ("also include Amsterdam", "prefer startup backgrounds").
* Start a new search for a different role.
Each channel is saved in the sidebar and can be revisited anytime.
What's next [#whats-next]
* [Key Terms](/en/docs/meet-mira/key-concepts): Agent, candidate pool, and other terms explained.
* [Quick Start](/en/docs/get-started/quick-start): Run your first sourcing task.
---
# Key Terms
Source: https://docs.mira.day/en/docs/meet-mira/key-concepts
> Plain-language definitions of the terms you'll see in MiraDay: Agent, Agent Team, teammate, channel, Skill, Ideal Candidate Profile, candidate pool, Qualifications, MRE, Mira Browser Extension, and Credit.
Here's what the key terms mean when you see them in MiraDay.
***
Agent [#agent]
Mira, MiraDay's AI agent, executes tasks on your behalf: understanding requirements, searching, evaluating, and delivering results. Mira works independently and brings you in at key decisions.
Agent Team [#agent-team]
The teammates Mira brings in to work a task with it, each owning a part. Every candidate search runs with at least one sourcing teammate, and hard ones get several. For other tasks, Mira assembles a team when it helps, or you can turn it on yourself. See [Agent Team](/en/docs/meet-mira/agent-teams).
Teammate [#teammate]
A member of an Agent Team. Mira is the team lead; teammates are the agents Mira brings in, each owning part of a task, such as people search (Scout), general research (Rowan), or fact-checking (Theo).
Channel [#channel]
A channel is the space for one task, a conversation with Mira. Most often it's a sourcing request: you describe a role and Mira delivers candidates, but a channel can be any request you give Mira. The channel is also where an Agent Team divides the work and collaborates on it. Channels are saved in the sidebar, so you can come back anytime, pick up where you left off, ask for more candidates, or adjust your criteria.
Skill [#skill]
A reusable workflow. Type / in the channel input to invoke a Skill instead of describing the same requirements from scratch; the Skill name appears as a chip in your message. Mira ships with built-in Skills, and you can create your own through conversation with skill-creator, import them as files, or manage them in your Skills settings.
Ideal Candidate Profile [#ideal-candidate-profile]
A structured summary of your search that Mira proposes before searching, in three lists: Must-Have, Nice-to-Have, and Avoid. You review and edit it, and the search runs against it once you confirm.
Candidate pool [#candidate-pool]
What you get back from a search: a curated pool of candidates who match your requirements. Each person comes with a name, current role, work experience, education, a Qualifications assessment of why they're relevant, and contact details.
In classic recruiting, a shortlist means the vetted final list of roughly 3 to 6 candidates left after outreach and screening. Mira's output is the broader, reasoned, sourced set, so we call it a candidate pool, not a shortlist.
Qualifications [#qualifications]
Mira's per-candidate assessment for your search: it checks the candidate against your must-haves one by one, showing where they meet, where the evidence is uncertain, and where they fall short, with the evidence behind each, plus an overall label: Good Match, Partial Match, or Low Match. It's specific to your search, not a generic bio.
MRE (Mira Reasoning Embedding) [#mre-mira-reasoning-embedding]
MRE is the reasoning-based retrieval model behind Mira's sourcing. Most search tools match on surface similarity: they can tell that two profiles look related, but not why. MRE reasons first. When MRE reads your requirement, it works out what you're really asking for and which underlying signals to look for, then searches on that reasoning rather than on keywords or look-alikes. That's why you can describe a role in plain language: a request like "has taken a product from 0 to 1" becomes a search for the signals behind it, such as founding-team roles and early-stage company history.
Learn more in [How MiraDay Matches Candidates](/en/docs/match-quality-and-trust/understanding-ai-matching).
Mira Browser Extension [#mira-browser-extension]
The Mira Browser Extension browses the web on your behalf, through your own logged-in browser session. It can reach LinkedIn, company websites, your ATS, and other platforms you're signed into to search for candidates and gather information. This is user-driven browser automation, not a platform-level integration. Useful when you need data that's not in MiraDay's built-in sources. See [Browser Extension](/en/docs/tools-and-teamwork/mira-browser-extension).
Credit [#credit]
The unit you "spend" when using MiraDay. Searches and contact reveals consume credits. You can see how many a channel used in its details.
---
# What is MiraDay
Source: https://docs.mira.day/en/docs/meet-mira/what-is-mira
> MiraDay is a Talent Intelligence platform for professional and leadership hiring. Its lead agent, Mira, reasons about who actually fits and delivers a candidate pool backed by evidence.
MiraDay is a Talent Intelligence platform for professional and leadership hiring. It reasons about who actually fits a role, not just whose profile uses the right words. [Why MiraDay](/en/docs/meet-mira/why-mira) explains the difference.
Its lead agent, Mira, works alongside you like a recruiting team. Every candidate search runs as an Agent Team: Mira hands the search to a sourcing teammate, brings in more teammates when the job calls for it, and coordinates the work. Whether a candidate moves forward stays your call.
***
How MiraDay works with you [#how-miraday-works-with-you]
Describe the role in plain language, no forms or boolean strings. Mira confirms an Ideal Candidate Profile with you, then searches across many sources, follows and cross-checks the signals over several steps, and returns candidates ordered by relevance, each with source-backed evidence for what's met, what's uncertain, and what's missing. [How MiraDay Works](/en/docs/meet-mira/how-mira-works) walks through the full loop.
What you can do today [#what-you-can-do-today]
Right now, MiraDay's primary use case is talent sourcing:
* **Describe a role in your own words, or upload a job description:** including implicit criteria like "startup DNA" or "has built something from scratch."
* **Get a candidate pool** with profiles, a Qualifications assessment, and recent work experience. 40 candidates by default, up to 100 if you ask.
* **Refine through conversation:** "more startup backgrounds" or "include Amsterdam too."
* **Reveal contact details:** email and phone on demand (uses credits).
* **Download structured data:** import into your ATS or spreadsheet, or share with your hiring manager or client.
* **Go further** with [Agent Team](/en/docs/meet-mira/agent-teams), [Skills](/en/docs/tools-and-teamwork/skills), and the [Mira Browser Extension](/en/docs/tools-and-teamwork/mira-browser-extension).
Where do the candidates come from? [#where-do-the-candidates-come-from]
| Source | What it covers |
| -------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Public talent data** | Aggregated from multiple public talent data sources and providers, including LinkedIn, X, GitHub, and Hugging Face, and data providers like Exa, Apollo, and RocketReach. Covers industries, regions, and seniority levels globally |
| **MiraDay's own index** | Continuously updated index of public professional profiles and job market data |
| **Mira Browser Extension** | Navigates recruiting platforms, company websites, forums, and internal systems on your behalf |
Who uses MiraDay [#who-uses-miraday]
MiraDay is built for professional and leadership hiring: in-house Talent Acquisition and executive/leadership search teams, and the executive-search and professional recruitment agencies that serve them, on the specialized, business-critical roles where evidence and judgment matter more than application volume.
Get started [#get-started]
* [Create Your Account](/en/docs/get-started/create-your-account): sign up and get started.
* [Quick Start](/en/docs/get-started/quick-start): run your first sourcing task.
---
# Why MiraDay
Source: https://docs.mira.day/en/docs/meet-mira/why-mira
> Keyword search matched words; semantic search matched meaning. Neither is the same as finding the right candidate. This is the problem MiraDay was built to solve.
Recruiting comes down to two things: volume and judgment. There are more places to look than ever, and the person who actually fits is rarely the one with the tidiest profile. MiraDay was built to handle the volume so you can focus on the judgment, on the professional and leadership roles where finding the right person matters most.
Matching was never the same as finding [#matching-was-never-the-same-as-finding]
* **Keyword search matches words.** Boolean strings find the profiles that happen to use your terms, and miss the strong candidate who described the same work differently, or never updated their title.
* **Semantic search matches meaning.** It returns people who read as similar on paper, but "similar" is not the same as "right for this role."
Both moved search forward. Neither answers the question a recruiter actually asks: who is the right person for this job, and why?
Mira reasons its way to the right candidate [#mira-reasons-its-way-to-the-right-candidate]
Instead of matching strings or look-alikes, Mira reasons across the signals the way a seasoned recruiter would: career trajectory, the work behind a title, company context, public work. Mira weighs each candidate against your must-haves and dealbreakers, surfaces the non-obvious candidates, and brings back the people who actually fit, with the reason behind every match. The reasoning-based retrieval behind Mira's sourcing, MRE, ranks first overall on BRIGHT, a public benchmark for reasoning retrieval, as of July 2026. See [How MiraDay Matches Candidates](/en/docs/match-quality-and-trust/understanding-ai-matching).
You stay the one who decides [#you-stay-the-one-who-decides]
Mira runs every search with an Agent Team: its teammates search across multiple sources, cross-check the signals, and rank by evidence, and a hard search gets more teammates working in parallel. Then Mira pauses at the decisions that are yours. You set the goal and make the calls; you don't manage each step. And Mira doesn't start from a blank page: top-agency recruiting playbooks come built in as [Skills](/en/docs/tools-and-teamwork/skills).
What changes for you [#what-changes-for-you]
* **From boolean strings to plain language.** Describe a role the way you would brief a colleague; Mira does the rest.
* **From sifting to deciding.** Mira reads the volume so your time goes to judgment, not scrolling.
* **From good-looking résumés to candidate pools you can trust.** Every candidate comes with the evidence, so the pool you put forward holds up with a hiring manager.
New here? Start with [What is MiraDay](/en/docs/meet-mira/what-is-mira), then [How MiraDay Works](/en/docs/meet-mira/how-mira-works).
---
# How MiraDay Searches
Source: https://docs.mira.day/en/docs/sourcing/agentic-search
> Describe the person you want in plain language. Mira searches across many talent sources and the open web, digs in step by step following the signals it surfaces, and returns candidates with the reason each one matches.
Agentic Search is how Mira finds people, and it works differently from LinkedIn Recruiter and traditional keyword search. Describe who you're looking for in plain language, and Mira works like a seasoned recruiter.
***
What it solves [#what-it-solves]
Traditional candidate search has two built-in limits:
* **One database.** The right people are scattered across professional networks, code communities, research platforms, and company sites. A single database is bound to miss some.
* **One pass.** A single keyword match tends to return people with good-looking résumés, not people who are actually a good fit.
Agentic Search is different on both counts: it pulls from **many sources** and advances in **multiple steps**, deciding where to look next from what the last step surfaced, until it finds the right people, including passive candidates who aren't actively looking for a job.
How it works [#how-it-works]
1. **You describe the need**, in everyday language: the role, the must-haves, and the nice-to-haves.
2. **A Sourcing Agent searches** across many talent sources and the open web, following signals (companies, tech stacks, projects, public work) and cross-checking them.
3. **You get explainable results**: a candidate pool where every person comes with the reason they match.
4. **You refine.** If the direction is off, just say so or add a constraint, and Mira adjusts.
**Want a specific recruiting platform?** Name it. Say you want to search only that platform and Mira keeps the search there; mention it without "only" and Mira gives it priority. If a platform you asked for exclusively can't be reached, Mira asks before switching to another source instead of switching on its own. Platforms that need your sign-in are reached through the [Mira Browser Extension](/en/docs/tools-and-teamwork/mira-browser-extension).
Core capabilities [#core-capabilities]
* **Multi-source, multi-step**: results from every source are merged and deduplicated, and each hop is cross-checked. Powered by [MRE](/en/docs/match-quality-and-trust/understanding-ai-matching).
* **Evidence-weighed matching**: Mira evaluates each candidate against your must-haves and dealbreakers and ranks them by the evidence, surfacing the non-obvious (like the strong engineer who never updated their job title). Every candidate carries the reason behind the match.
* **On-demand contact**: once the right people are found, reveal email or phone as needed (each reveal uses credits).
Describe the role well [#describe-the-role-well]
Mira understands natural language, but more context gives it more to reason about. A weak prompt and a strong one return very different candidate pools:
| Weak | Strong | Why it matters |
| ------------------------- | ----------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------- |
| "Find me a developer" | "Senior backend engineer, 5+ years Python, has built microservices at scale, preferably from fintech" | More context means more signals to match against |
| "Marketing person in NYC" | "B2B SaaS marketing manager in NYC, has run product launches, comfortable with data-driven campaigns" | Concrete signals (launches, data-driven campaigns) give Mira something to check against, not just a title and a city |
| "We need someone good" | "Someone who's taken a product from 0 to 1, ideally at a Series A to B startup" | Mira reads "0 to 1" as early-stage building experience |
You can also lean on the implicit criteria you'd say out loud. Mira reasons about what they mean rather than matching them as keywords:
| You say | Mira reads it as |
| -------------------------------- | ----------------------------------------------------------------------- |
| "0 to 1" or "built from scratch" | early-stage company building, taking something from nothing to launch |
| "startup DNA" or "scrappy" | operates independently, comfortable with ambiguity and little structure |
| "high-growth background" | inferred from company size, funding history, and growth trajectory |
| "5+ years of experience" | read from the real career timeline, not a profile tag |
| "has built a team" | hired and led a team, not just worked on one |
What's next [#whats-next]
* [Must-haves & Dealbreakers](/en/docs/sourcing/search-filters-and-criteria): how Mira reads your requirements and builds the Ideal Candidate Profile.
* [How MiraDay Matches Candidates](/en/docs/match-quality-and-trust/understanding-ai-matching): how MRE reasons to surface the right candidates.
* [Agent Team](/en/docs/meet-mira/agent-teams): put a team on a complex search.
---
# Inside a Candidate Profile
Source: https://docs.mira.day/en/docs/sourcing/candidate-profile
> How to read a candidate in MiraDay. The card fields, the Qualifications assessment and fit label, the full profile view, and how to evaluate fit.
Every candidate in your pool comes with a profile built for your search. Each profile is a single person: Mira merges the same person found across multiple public sources into one candidate, so you won't see duplicates, and the profile keeps filling in as more is gathered from those sources. Here's how to read it and judge fit.
What's on a candidate card [#whats-on-a-candidate-card]
In the results panel, each candidate shows as a card with:
* **Header**: name, and clickable source links to the person's profiles on the sources the search drew on, such as LinkedIn, GitHub, Hugging Face, and X (Twitter), plus the recruiting platforms they were found on. Ask Mira to search a source like GitHub or Hugging Face, and links to those profiles appear on the card.
* **Current role** (title · company · dates), then **location · highest education** (degree, school, field, years) on the line below.
* A **Qualifications** preview with the overall fit label, showing the first five requirements, with **Show more** to expand the rest.
Click a card to open the full **Profile**. It adds Email and Phone rows, an **About** summary, the full per-requirement assessment, complete work experience and education, skills and languages, GitHub and Hugging Face activity, papers, awards, certifications, and courses where available, and a **More** section with web and news mentions (article title, publish date, and a relevant line). Use the arrows at the top to move to the previous or next candidate.
The results panel opens in **Card view**. Switch to **List view** with the toggle at the top of the panel to see a compact table (candidate and fit label, links, email, phone, location, job title, company, school, degree, major), useful for scanning many at once.
The fit label [#the-fit-label]
Each candidate carries a color-coded fit label that sums up how well they match your search:
* **Good Match** (green): the strongest matches against your requirements.
* **Partial Match** (yellow): partial matches worth a look.
* **Low Match** (grey): the furthest-out candidates, weakest against your requirements.
Candidates are ordered by fit, so the strongest matches sit at the top.
The Qualifications assessment [#the-qualifications-assessment]
The **Qualifications** section lists your must-haves one at a time. For each, Mira marks whether the candidate meets it (check), the evidence is thin (dash), or they fall short (cross), with the evidence behind each. It's written for your specific search, not a generic bio.
How to evaluate a candidate [#how-to-evaluate-a-candidate]
The label and Qualifications are a fast read. Before you decide, look deeper:
* **Look beyond titles.** The same title means different things at different companies. "Senior Engineer" at a 10-person startup is a very different role from the same title at a large company. Use company context to read the real scope.
* **Check the career trajectory.** Progression speed and the types of companies across roles say a lot. Someone who moved enterprise, then startup, then growth-stage has a different profile from someone who stayed in one setting.
* **Use the Qualifications as a starting point.** It highlights what's most relevant to your search. Read the full work history and education for the nuance it can't capture.
Contacting candidates [#contacting-candidates]
Each card has an Email and a Phone row. Click **Get email** or **Get phone** to reveal it, on the card, in List view, or in the full profile. While a phone lookup runs, the row shows **Getting phone…**. To reveal several at once, hover a card's avatar and tick its checkbox (or use **Select all**), then click **Get emails** or **Get phones** in the bar at the bottom of the panel, up to 50 people per run. Revealing uses credits, and a phone costs more than an email.
Save a candidate [#save-a-candidate]
To keep someone beyond this search, click the **Add to contacts** icon at the top right of their card (or the button in their full profile). To save several, tick their checkboxes and click **Add to contacts** in the bottom bar. They're added to [Contacts](/en/docs/tools-and-teamwork/contacts), your workspace-level list, where you can enrich them and keep them up to date.
What's next [#whats-next]
* [Refining Your Search](/en/docs/sourcing/refine-and-export): sharpen the pool by conversation.
* [Contacts](/en/docs/tools-and-teamwork/contacts): keep and organize the people you save.
* [Export & Share](/en/docs/get-started/export-and-share): reveal contacts and export the pool.
---
# Refining Your Search
Source: https://docs.mira.day/en/docs/sourcing/refine-and-export
> Sharpen your candidate pool by talking to Mira: ask for more, adjust criteria, or change direction in the same conversation. Then export the candidates you want.
A first search gets you a strong starting pool, rarely the final one. The fastest way to sharpen it is to keep talking to Mira in the same conversation, the way a recruiter recalibrates the profile after seeing the first batch or hearing feedback from the hiring manager or client.
Refine by conversation [#refine-by-conversation]
After a pool comes back, tell Mira what to change. Each follow-up continues the same channel and adjusts the search.
| To do this | Say something like |
| ----------------------- | ---------------------------------------------------- |
| **Get more candidates** | "Find me 10 more" or "Keep searching" |
| **Adjust criteria** | "Focus on candidates with more startup experience" |
| **Change location** | "Only London, not remote" or "Include Amsterdam too" |
| **Broaden the search** | "Also include mid-level candidates" |
| **Narrow the search** | "Only people currently at high-growth startups" |
You don't start over. Mira keeps the context of the search and applies your change on top of it.
To change specific criteria yourself, expand the **Ideal Candidate Profile** bar at the top of the results panel, edit the rows, and click **Search**. Mira starts a new round with the updated profile. If a search is still running, you can **Wait** for it or **Start new search**. If the profile gets updated in the conversation before you send your edits, those unsent edits are discarded. See [Must-haves & Dealbreakers](/en/docs/sourcing/search-filters-and-criteria).
Choosing how many [#choosing-how-many]
Mira returns 40 candidates by default. To change that, say a number when you describe the role ("find me 50 ML engineers"), or ask for more at any point in the conversation.
Start a new search [#start-a-new-search]
For a different role, click **New channel** in the sidebar. Each conversation is saved there as a channel, so you can jump back to any earlier search and pick up where you left off.
Export and share [#export-and-share]
Once the pool looks right, click **Download profiles** at the top of the results panel for an Excel (.xlsx) file, or tick candidates and click **Bulk download** to export just those. For what the file contains and how to share it, see [Export & Share](/en/docs/get-started/export-and-share).
What's next [#whats-next]
* [Inside a Candidate Profile](/en/docs/sourcing/candidate-profile): how to read the candidates you get back.
* [Export & Share](/en/docs/get-started/export-and-share): reveal contacts and export the pool.
* [Agent Team](/en/docs/meet-mira/agent-teams): put a team on a hard or long-running search.
---
# Must-haves & Dealbreakers
Source: https://docs.mira.day/en/docs/sourcing/search-filters-and-criteria
> How Mira turns your requirements into an Ideal Candidate Profile. What goes in Must-Have, Nice-to-Have, and Avoid, and how each one shapes the search and the ranking.
Before it searches, Mira turns your brief into an **Ideal Candidate Profile**: a structured summary you review and confirm. It's three lists, and each one steers the search differently. Getting your requirements into the right list is how you control the results.
The three lists [#the-three-lists]
| List | What goes here | How it shapes the search |
| ---------------- | ------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Must-Have** | The non-negotiables a candidate has to meet | The hard bar. A candidate who clearly fails one is filtered out. Rows phrased as a preference ("ideally...") are treated as preferences, and criteria public profiles rarely show, like salary expectations, don't rule anyone out when the information is missing. |
| **Nice-to-Have** | Preferences that make someone stronger, but won't rule them out | Broadens the search and lifts matching candidates up. It never excludes anyone. |
| **Avoid** | Dealbreakers: a competitor to steer clear of, an agency or consulting-only background | Not used to search. Mira applies it after retrieving candidates, to filter out or downrank anyone who matches. |
The lists are independent, and the same criterion shouldn't sit in two of them. Any list can be empty, as long as one isn't.
Keep protected characteristics out of your criteria. Age, gender, race, religion, nationality, disability, and similar traits aren't valid requirements: Mira is designed not to filter or rank on them, and in most markets using them in hiring criteria is unlawful.
Hard filters vs reasoned criteria [#hard-filters-vs-reasoned-criteria]
You don't have to sort your requirements into "filters" and "fuzzy" yourself. Inside **Must-Have**, Mira reads each row and decides how to use it:
* **Literal filters** for things that map to a field: a title, a seniority, a location, a company.
* **Reasoned criteria** for the rest, the implicit ones like "took a product from 0 to 1" or "high-growth background", which Mira interprets rather than keyword-matches (see [How MiraDay Searches](/en/docs/sourcing/agentic-search)).
So you just state the requirement in plain language. Mira handles whether it's a hard filter or something to reason about.
Review, edit, and confirm [#review-edit-and-confirm]
Mira proposes the Ideal Candidate Profile on a card before it searches. You're in charge of it:
* **Click Edit** to change any list: add a **Must-Have**, delete a row and re-add it under **Nice-to-Have**, or add a competitor to **Avoid**. While editing, **Undo** and **Redo** step through your changes.
* When it looks right, click **Search**. Whether you took it as-is or edited it first, the search runs against the profile on the card. Criteria you delete stay out of the search.
Change the profile after a search [#change-the-profile-after-a-search]
Once results come back, the profile stays at the top of the results panel as a collapsed **Ideal Candidate Profile** bar that shows how many must-have, nice-to-have, and avoid criteria it holds. Expand it to add, edit, or remove criteria one by one, then click **Search** (it turns active once you've changed something). Mira starts a new search in the conversation with the updated profile; the earlier results and their profile stay as they were. If the previous search is still running, Mira asks whether to **Wait** for it to finish or **Start new search** right away. You can also keep adjusting the profile by telling Mira in the conversation.
Tips [#tips]
* **Put non-negotiables in Must-Have, not Nice-to-Have.** A requirement in Nice-to-Have won't gate anyone out; it only boosts.
* **Use Avoid for what rules a candidate out**, like a specific competitor or an off-target industry, rather than burying it in the description.
* **Keep each row to one atomic criterion**, so Mira can weigh them cleanly.
What's next [#whats-next]
* [How MiraDay Searches](/en/docs/sourcing/agentic-search): how the confirmed profile drives a multi-source search.
* [Inside a Candidate Profile](/en/docs/sourcing/candidate-profile): how to read the candidates that come back.
* [How MiraDay Matches Candidates](/en/docs/match-quality-and-trust/understanding-ai-matching): how MRE reasons about your criteria.
---
# Tips & Best Practices
Source: https://docs.mira.day/en/docs/sourcing/tips-and-best-practices
> Practical advice for getting the most out of Mira's sourcing capabilities.
How should I describe a role to Mira? [#how-should-i-describe-a-role-to-mira]
Don't think of it as filling out a form. Describe the role the way you'd explain it to a fellow recruiter:
> "I need a product manager who's done B2B SaaS, ideally at a Series A-B company. Someone who can talk to customers and also go deep on data. Based in London or willing to relocate."
This natural briefing style gives Mira more context to work with than a formal JD with bullet points.
What are implicit criteria, and why do they matter? [#what-are-implicit-criteria-and-why-do-they-matter]
The criteria you'd normally keep in your head or tell a recruiter verbally are exactly what makes Mira different from keyword search:
* "Someone who's been through hyper-growth."
* "Has experience in a regulated industry."
* "Can operate independently without a big team around them."
* "The kind of person who'd thrive in a chaotic early-stage environment."
Mira Reasoning Embedding surfaces candidates whose background fits these implicit patterns.
Should my first search be perfect, or can I refine? [#should-my-first-search-be-perfect-or-can-i-refine]
Refine. Your first search doesn't need to be perfect. A common workflow:
1. **Round 1**: Describe the core requirements, keep it open.
2. **Review**: Look at your candidate pool, notice patterns in who's strong and who's not.
3. **Round 2**: "More candidates with payment industry experience, like Jane Doe has." Anchor on the skills you liked, not the person.
4. **Round 3**: Fine-tune geography, seniority, or company type.
Each round builds on the conversation context. Mira remembers everything.
Can my team collaborate on a candidate pool? [#can-my-team-collaborate-on-a-candidate-pool]
Yes. Download the candidate pool data file and share it with:
* **Hiring managers**: Let them review candidates before you reach out.
* **Team members**: Split candidate outreach across the team.
* **Your ATS / tracking sheet**: Import for tracking and talent pipeline management.
Should I check LinkedIn after seeing the Qualifications assessment? [#should-i-check-linkedin-after-seeing-the-qualifications-assessment]
Yes. The Qualifications assessment gives you a quick read, but LinkedIn profiles often contain additional signals:
* Posts and articles they've written (thought leadership).
* Recommendations from colleagues (peer validation).
* Activity patterns (are they actively looking?)
* Mutual connections (warm introduction potential).
Should I run multiple roles in one channel? [#should-i-run-multiple-roles-in-one-channel]
No. Create separate channels for different roles. This keeps conversations focused and makes it easier to revisit searches later. Don't mix "Find me a backend engineer" and "Also find a designer" in the same channel.
Can I revisit a channel later? [#can-i-revisit-a-channel-later]
Yes. Channels stay in your sidebar until you delete them. Come back tomorrow and say "Any new candidates matching these criteria?" or "Let's try a different angle, focus on candidates who are currently at consulting firms." Mira picks up right where you left off.
---
# Voice & Audio Input
Source: https://docs.mira.day/en/docs/sourcing/voice-input
> Speak your search and Mira transcribes it into the channel input, or upload an audio file and Mira reads it to understand your requirements.
Voice input lets you describe a candidate search out loud instead of typing it. Mira transcribes what you say into the channel input, so you can review or edit before sending. You can also upload an audio file you already have, and Mira reads it to understand your requirements and gets to work.
***
What it's good for [#what-its-good-for]
* Long requirements, where talking is faster than typing every detail.
* Briefing the way you'd brief a colleague. The natural phrasing usually carries more context than a tidy bullet list.
* Moments when typing is awkward but you don't want to lose the thought.
How to use it [#how-to-use-it]
1. Open a channel and find the microphone button in the channel input.
2. Click the microphone button and start speaking.
3. When you're done, click the confirm button. Mira transcribes the recording and adds the text to the channel input, after anything you've already typed.
4. Review or edit the transcription, then send it as usual.
If you change your mind, click the cancel button to discard the recording. Nothing is sent until you choose to.
Each recording is capped at 5 minutes; when you reach the limit Mira stops automatically and transcribes what it has. For a very long brief, record it in shorter segments, or upload it as a file instead.
Upload an audio file [#upload-an-audio-file]
Already have a recording? Attach an audio file instead of speaking live: a meeting recording, an intake call with a hiring manager, or interview audio. Mira reads it the way it reads an uploaded JD, works out the requirements from it, and gets to work. Unlike live voice, the content isn't placed in your input box as editable text; Mira takes it in as part of the channel.
1. In the channel input, attach an audio file the same way you'd attach a JD or résumé.
2. Common formats are supported: MP3, M4A, WAV, WebM, and OGG. Each file can be up to 100 MB and up to 2 hours long.
3. Mira reads the audio, understands the requirements in it, and uses them in your channel.
Recording laws vary: make sure you have the consent required in your jurisdiction before recording and uploading calls or interviews.
Supported languages [#supported-languages]
Live voice input uses your MiraDay interface language as a hint for transcription, so it works best when you speak in that language. To dictate in another language, switch the interface language first. Uploaded audio files are transcribed in whatever language they're spoken in: Mira detects it automatically.
Privacy [#privacy]
* For live voice input, the recording is used only to produce the transcription that fills the channel input, and Mira does not store it.
* Uploaded audio files are stored as channel attachments, like any other file you upload.
* Transcription is handled by a third-party transcription provider.
* The transcribed text becomes part of your channel like any other input. See [Privacy & Compliance](/en/docs/match-quality-and-trust/privacy-policy) for how channel content is handled.
---
# Connectors
Source: https://docs.mira.day/en/docs/tools-and-teamwork/connectors
> Connect your Gmail or Outlook so Mira can draft, send, and summarize email as part of a channel, right where the work happens.
Connectors link your own accounts to MiraDay, so Mira can use them while it works in a channel. The one most recruiters reach for first is email: connect your **Gmail** or **Outlook**, and Mira can draft, send, and summarize email as part of a channel, without you switching back and forth to your inbox.
***
Connect your email [#connect-your-email]
1. Open **Connectors** from **Settings**, or from the connectors menu in the channel input.
2. Find **Gmail** or **Outlook** and click **Connect**.
3. **Authorize** Mira when prompted. You're sent to sign in and grant access, then brought back.
4. Once connected, your mailbox is ready to use in any channel.
Connected services are listed first. Click a connector to open its detail page, where you can turn each tool on or off; tools marked **Approval** only run after you confirm. **Disconnect** at any time (Mira asks you to confirm) and access is revoked right away; you can reconnect later.
What Mira can do with your email [#what-mira-can-do-with-your-email]
Once your mailbox is connected, ask Mira to work with email right inside a channel:
* **Draft.** Mira writes the email in the **Email Canvas**, with **To**, **Cc**, **Bcc**, **Subject**, and **Body** all editable inline, and you can add attachments. Review it and make any changes before it goes anywhere.
* **Send.** When the draft is ready, click **Send**. Mira asks you to confirm once more before anything leaves your mailbox, so nothing is sent until you approve it.
* **Forward.** Forward an email from within a channel. Its original attachments come along; forwarded emails can't take new attachments.
* **Read and summarize.** Point Mira at your inbox to pull in a long thread and get the gist, or catch up on what's come in, without leaving the channel.
Privacy and security [#privacy-and-security]
* Connectors use your own authorization. Mira uses a connected account only to do what you ask inside a channel.
* Nothing is sent without your confirmation, and you can disconnect a connector at any time from the Connectors page; its access is revoked right away.
* All data handling follows MiraDay's [Privacy Policy](/en/docs/match-quality-and-trust/privacy-policy).
What's next [#whats-next]
* [Agent Team](/en/docs/meet-mira/agent-teams): the teammates that do the work inside a channel.
* [Skills](/en/docs/tools-and-teamwork/skills): save a repeatable workflow and reuse it in any channel.
* [Files](/en/docs/tools-and-teamwork/files): everything your channels produce or use, in one place.
---
# Contacts
Source: https://docs.mira.day/en/docs/tools-and-teamwork/contacts
> Your saved people in one place. Add candidates and connections from a search or import your own network, then enrich each profile and reveal contact details, all from a workspace-level contact list.
Contacts is your saved people: candidates and connections you want to keep beyond a single channel. Add someone from a search, or import the contacts you already have, and enrich any profile when you need fresher details.
***
Open Contacts [#open-contacts]
Open **Contacts** from the sidebar. The page lists everyone you've saved, with a search box and a table you can scan.
Add people to Contacts [#add-people-to-contacts]
There are three ways someone lands in your list:
* **From a search.** On a candidate in your results, click **Add to contacts**. To save several at once, select the ones you want with their checkboxes and add them together.
* **Import your own files.** Click **Import** and drop in PDF, Word, Excel, CSV, or HTML files (up to 10 files, 10 MB each). Mira parses them (usually 5–30 seconds), shows you a preview of what's valid, duplicated, or failed, then imports in the background so you can keep working; the contacts that come through are ready to use right away.
* **From the browser extension.** Candidates Mira finds through the [Mira Browser Extension](/en/docs/tools-and-teamwork/mira-browser-extension) can be saved the same way.
The contact list [#the-contact-list]
The list shows each person as a row with name, job title, company, location, contact, profile links, and where they came from. **Search** by name, title, or company to find someone, and the **Source** label (User Import, Mira Search, or Mira Browser Extension) tells you how each contact got there.
Inside a contact [#inside-a-contact]
Click a contact to open their full profile:
* **Contact details.** Email and phone. Click **Get email** or **Get phone** to reveal them on demand (each reveal uses credits), then copy them.
* **Profile.** Experience, education, skills, languages, and, where the data exists, GitHub activity, publications, and social links.
* **Enrich.** Click **Enrich**, review the estimated credits, and click **Start**. Mira works in the background (usually 10–60 seconds), then shows **Review enrichment changes**: pick the updates you want and click **Apply selected**. A contact needs a LinkedIn URL or email before it can be enriched.
Notes [#notes]
Jot down anything you want to remember about a person, so your context on each contact stays in one place.
You can also **remove** a contact from your list at any time.
What's next [#whats-next]
* [Inside a Candidate Profile](/en/docs/sourcing/candidate-profile): how to read a candidate before you save them.
* [Export & Share](/en/docs/get-started/export-and-share): export the candidate data file from a search.
* [Connectors](/en/docs/tools-and-teamwork/connectors): connect Gmail or Outlook to work with email in a channel.
---
# Files
Source: https://docs.mira.day/en/docs/tools-and-teamwork/files
> Every file your channels produce or use, uploaded résumés and JDs plus the data files and reports Mira generates, gathered in one place across all your channels.
Files gathers what you upload and what Mira produces across all your channels, so you don't have to dig through each channel to find a résumé, JD, or report.
***
Open Files [#open-files]
Open **Files** from the sidebar. The page lists files from across all your channels, newest first.
Where files come from [#where-files-come-from]
Use the tabs at the top to filter by source:
| Tab | What it shows |
| ------------------ | -------------------------------------------------------------------------------- |
| **All** | Every file, from every channel |
| **Mira generated** | Files Mira produced: candidate data files, reports, and other outputs |
| **My uploads** | Files you uploaded into a channel, such as a JD, a résumé, or an audio recording |
Files shared with you carry a **Shared** label on the file row.
Find a file [#find-a-file]
When the list grows, narrow it down:
* **Search** by file name.
* **Filter** by type, channel, or date created.
* **Group** by channel or by file type, or leave it ungrouped.
* **Sort** by most recent, oldest, or file size.
Files cover the formats your channels work with: candidate data files (Search Result), documents, spreadsheets, presentations, PDFs, images, code, and plain text.
What you can do with a file [#what-you-can-do-with-a-file]
Click a file to preview it; **Reveal in channel** and **Download** are in the row's ⋯ menu.
* **Preview** a supported file inline without leaving the page.
* **Reveal in channel**: jump to the channel the file belongs to, in context.
* **Download** a single file, or select several and **Batch download** them together.
What's next [#whats-next]
* [Export & Share](/en/docs/get-started/export-and-share): the candidate data file Mira generates from a search.
* [Voice & Audio Input](/en/docs/sourcing/voice-input): upload a recording and Mira works out the requirements from it.
* [Connectors](/en/docs/tools-and-teamwork/connectors): connect Gmail or Outlook to work with email in a channel.
---
# Mira Browser Extension
Source: https://docs.mira.day/en/docs/tools-and-teamwork/mira-browser-extension
> Let Mira operate your own Chrome browser in a tab dedicated to the channel, reaching sites and sign-ins that aren't in MiraDay's built-in sources.
The Mira Browser Extension lets Mira operate your own Chrome browser to reach information that isn't in its built-in sources, like sites that need a sign-in. It works in a dedicated tab in your browser, using your own logins, so you can watch what it does and step in whenever you want.
***
What you can use it for [#what-you-can-use-it-for]
* **Browse a site you name.** Give Mira a specific website you want visited and it browses it for you, then brings back what it found.
* **Reach candidates behind a sign-in.** Open a recruiting site you're logged into and Mira reads profiles, then brings qualified candidates back for matching.
* **Research companies and teams.** Ask about a company's leadership, recent funding, or who's on a specific team, and Mira gathers it from the live web.
* **Reach other SaaS tools you're signed into.** Mira works through your own browser session, so it can reach your ATS, CRM, or other accounts you're already logged into.
Install and connect [#install-and-connect]
1. Find **Mira Browser Extension** on the [Chrome Web Store](https://chromewebstore.google.com/detail/mira-chrome-extension/eehnmnpglnfhkkcepfdlnnolobnjfcoh) and click **Add to Chrome**. You can also install it from **Settings**, then **Browser Extension**, with the **Install extension** button (desktop only).
2. A small extension icon appears in the channel input (desktop only; it doesn't show on mobile). While Mira is detecting the extension, the button is disabled. If the extension isn't installed, clicking it opens install steps; if it's installed but not connected, clicking it offers **Connect**, **Authorize**, **Update**, or **Upgrade**, depending on what's needed; once connected, it works as a toggle (see below).
3. The first time Mira needs your browser, it shows a **Mira requests control of your browser** card that lists what you're authorizing: Mira controls your browser, including sites you're logged into; it only operates in the channel's dedicated tab, not your other tabs; all actions are logged and viewable; and sensitive actions like payment or deletion ask you first. Click **Authorize**; it's remembered after that, so there's nothing to set up per channel.
If the extension isn't ready when a channel needs it, Mira shows a card that walks you through installing, authorizing, or reconnecting, then picks the channel back up on its own once you're connected.
If your extension is older than the version Mira requires, browsing is blocked until you update: Mira shows an **Update the Mira browser extension** card, marked **Update required**, with no option to skip. On more recent versions of the extension, click **Update now** and the update installs in one click, without leaving the page. Otherwise, or if the one-click update doesn't go through, update by hand:
1. Go to `chrome://extensions/`, turn on **Developer mode** in the top right, then click **Update** in the top left.
2. If it still won't update, find the Mira extension and click **Remove**, then use the **Open store** button on the card and click **Add to Chrome** to reinstall.
Once you've updated, Mira picks the channel back up.
Using it in a channel [#using-it-in-a-channel]
Just type what you need done in your channel. Mira decides on its own when browsing is the best way to get what you need, then works in a dedicated tab in your own browser. When the extension is connected, its button also works as an optional toggle for the current turn: leave it off and Mira decides, or turn it on to have Mira prefer the browser this time.
If Mira hits a step that needs you, like signing in with a username and password, a CAPTCHA, or a two-factor prompt, it **pauses and shows a Continue button** and waits. You can also take over the tab yourself at any point; Mira waits and resumes when you click **Continue**. You can open **View Browser Session** to watch it work, and sensitive actions, like a payment, a deletion, or sending something, always ask for your confirmation first.
Works with your team [#works-with-your-team]
Teammates in an Agent Team can use the extension to browse too. You grant browser access once, up front, and the teammates Mira spins up for the channel can use the extension the same way, in your browser.
Platform compliance [#platform-compliance]
Many recruiting platforms have terms of service that restrict automated access. The Extension is built to work within reasonable usage patterns, but responsibility for complying with third-party platform terms rests with you. We recommend using it for targeted, specific searches rather than bulk scraping, and keeping your own platform accounts in good standing.
Privacy and security [#privacy-and-security]
* Mira uses your own browser and your own logins, only in the channel's dedicated tab, not your other tabs.
* Mira doesn't store your passwords or cookies; it works through the session you're already signed into.
* Sensitive actions ask for your confirmation first, and every action is logged and viewable.
* The Extension works on Chrome only.
* All data handling follows MiraDay's [Privacy Policy](/en/docs/match-quality-and-trust/privacy-policy).
What's next [#whats-next]
* [Agent Team](/en/docs/meet-mira/agent-teams): the teammates that share the browser during a channel.
* [Connectors](/en/docs/tools-and-teamwork/connectors): connect Gmail or Outlook to work with email in a channel.
* [How MiraDay Searches](/en/docs/sourcing/agentic-search): the built-in sources Mira searches before it reaches for the browser.
---
# Skills
Source: https://docs.mira.day/en/docs/tools-and-teamwork/skills
> Save your recurring workflows as reusable Skills and invoke them with / in the channel input. Built-in recruiting Skills come from the Mira team; your own are yours to create, edit, and share.
Finding the right candidate is a win. Capturing how you did it makes it repeatable, and that's a Skill. Save the workflows you use again and again (a sourcing brief format, an evaluation rubric, a reporting structure) and reuse them in any channel instead of describing the same requirements from scratch each time.
***
What are Skills [#what-are-skills]
A Skill is a set of saved instructions that tells Mira how to handle a recurring job: your intake questions, an evaluation rubric, a client-ready write-up format, even which tools to run and in what order. Mira runs the job end to end rather than treating it as a one-off prompt. Skills come in two kinds:
* **Official Skills**: built and maintained by the Mira team, listed in the **Official** tab. More on these below.
* **Personal Skills**: created by you, visible only to your account.
Export a Skill and share the file so your whole team works the same way.
Official Skills [#official-skills]
Official Skills appear in the **Official** tab. You invoke them like any other Skill, or Mira applies one automatically when your request matches it (see [Invoke a Skill](#invoke-a-skill)).
skill-creator [#skill-creator]
The one official Skill you'll reach for again and again is **skill-creator**: it builds and edits your own Skills through conversation, with no files to write.
* **What it does:** describe a workflow in plain language, and skill-creator turns it into a Skill, drafting the instructions for you.
* **How to use it:** in your Skills settings click **New Skill**, then **Create via Chat** (Mira opens a channel running skill-creator), or type **/skill-creator** in any channel. See [Create a Skill](#create-a-skill) for the full flow. To revise one of your Skills later, open it and click **Edit**.
* **When to use it:** when you have a repeatable process (an intake checklist, an evaluation rubric, or a client-ready report format) and want it saved as a reusable Skill without editing a `SKILL.md` by hand.
Invoke a Skill [#invoke-a-skill]
1. Click the **Skills** button in the channel input toolbar, or type **/**. A skill picker appears with your available Skills; **View all skills** takes you to your Skills settings.
2. Pick a Skill, or keep typing to filter by name.
3. Describe the specifics of this run and send. The Skill name appears as a chip in your message, and Mira follows the Skill's instructions for the channel.
Mira can also apply a relevant Skill automatically: every Skill carries a short description, and when your request matches it, Mira routes to that Skill on its own.
Create a Skill [#create-a-skill]
Open your Skills settings (**Settings**, then **Skills**) and click **New Skill**. There are two ways:
**Create via Chat.** Mira opens a new channel with the official **skill-creator** Skill, which turns the workflow you describe into a Skill through conversation. When the draft is ready, Mira delivers it as a card in the conversation; preview it and click **Add to my skills** to save it.
**Import File.** Drop in a `.md`, `.skill`, or `.zip` file (up to 10 MB). A single `.md` file must be named `SKILL.md`. Mira parses the `SKILL.md` frontmatter, adds the Skill, and opens its detail page showing the name, trigger, source, and files. A Skill package contains:
* `SKILL.md`, required: frontmatter with the Skill ID (slug) and a short description (up to 1024 characters; Mira uses it to decide when to auto-apply the Skill), followed by the instructions in markdown (up to 100 KB).
* Optional `scripts/`, `references/`, and `assets/` folders for supporting files.
If the name matches an official Skill, rename it in the `SKILL.md` frontmatter and upload again. If it matches one of your own Skills, Mira asks whether to replace it; click **Replace and save** to overwrite.
The Skill detail page [#the-skill-detail-page]
Open any Skill in your Skills settings to see its detail page: the description, its source, how it triggers (such as **Auto Routing**), and an on/off toggle. Your personal Skills also show their creation or last-edit time, how they were created (upload or chat), their version, and the Skill's files (tap a row for a read-only preview). For your personal Skills you can also:
* **Edit**: hand the change to **skill-creator**. Settings close and Mira continues in the channel where the Skill was created, or opens a new channel if that one is gone or the Skill was imported from a file, and applies the change for you.
* **Export .skill** (in the ⋯ menu): download the Skill as a `.skill` file.
* **Delete** (in the ⋯ menu): remove it permanently. This removes `SKILL.md` and everything under `scripts/`, `references/`, and `assets/`, and can't be undone.
Manage your Skills [#manage-your-skills]
Your Skills settings list everything in three tabs, **All**, **Official**, and **Personal**.
**Enable or disable** any Skill with its toggle; disabled Skills aren't used. Editing, exporting, and deleting happen on the Skill's detail page.
FAQ [#faq]
What official Skills are available? [#what-official-skills-are-available]
The **Official** tab lists the Skills the Mira team provides and has enabled for your account, including **skill-creator**, which helps you create and edit your own Skills through conversation. New official Skills appear there as the Mira team releases them.
Can I share a Skill with teammates? [#can-i-share-a-skill-with-teammates]
There's no built-in sharing yet. Export a Skill as a `.skill` file, and anyone can import it from their Skills settings via **New Skill** and **Import File**. This is how a personal Skill gets passed on and used across the whole team.
What's next [#whats-next]
* [How MiraDay Searches](/en/docs/sourcing/agentic-search): the search flow that Skills plug into.
* [Voice & Audio Input](/en/docs/sourcing/voice-input): another fast way to describe what you need.
---
# Hiring Doesn't Have an Efficiency Problem. It Has a Recognition Problem.
Source: https://blog.mira.day/en/blog/hiring-recognition-problem
> The right candidates are often already in your pipeline; they just don't get recognized. The real bottleneck isn't speed, it's judgment.
For years, the hiring industry has been obsessed with efficiency.
Faster pipelines.
More automation.
Better sourcing tools.
And yet, great candidates are still being missed every day.
Not because they don't exist.
But because they're not recognized.
The biggest myth in hiring [#the-biggest-myth-in-hiring]
We like to believe that hiring is a supply problem.
> "We just need more candidates."
So we build larger funnels. We scrape more profiles. We optimize response rates.
The right candidates are often already in your pipeline. They've applied. They've been sourced. They've even been reviewed.
And still, they get passed over.
Why great candidates get missed [#why-great-candidates-get-missed]
Think about the last VP-level hire you actually closed. Odds are, the person didn't come from the top of any search result. Their LinkedIn probably looked ordinary: "Director of Engineering at some Series B SaaS." What made you pause was likely a pattern only you would spot: three roles in a row taking on bigger scope, a company choice that signaled real judgment, a short stint at a no-name startup that most filters would flag as a red flag but actually meant ownership.
Keyword search doesn't catch that. Filters don't either. But you picked up on it within ninety seconds of reading the profile, because you've seen that shape before.
That's the gap. It's not random. It's not "human error." It's a signal problem.
Recruiters don't evaluate candidates purely on keywords or resumes. They rely on pattern recognition built from past hires, contextual intuition about what "good" actually looks like for a specific team.
These signals are invisible to most systems.
So we end up building tools that optimize for what's easy to measure, not what actually matters.
Efficiency doesn't fix misjudgment [#efficiency-doesnt-fix-misjudgment]
Most hiring tools today promise the same thing:
> "We help you move faster."
But speed doesn't solve the real issue. If your judgment is off, moving faster just means you reject the right candidates more efficiently. That's not a hiring problem solved. That's a hiring problem scaled.
What if we could learn how great recruiters think? [#what-if-we-could-learn-how-great-recruiters-think]
Top recruiters don't just process candidates. They interpret them.
They read between the lines and catch what filters don't: the lateral move that was actually a calculated bet, the person who left a well-known name for a Series B nobody's heard of. And they get better over time: every hire, every miss, every client debrief adds to a body of judgment that's genuinely hard to replicate.
That's the capability we're trying to build into MiraDay.
From tools to Agents [#from-tools-to-agents]
What we're building is not another sourcing tool.
We're building an Agent that learns from how recruiters make decisions: what "good" has looked like on past searches, which signals reliably predict fit, which pattern of experience tends to get passed over unfairly.
The goal is that over time, it surfaces candidates the way a senior recruiter would after closing a hundred searches just like it, not from keyword overlap, but from the judgment that comes with having seen the pattern before.
We're in early stages. But that's the direction.
So the right people don't get missed [#so-the-right-people-dont-get-missed]
Hiring will never be perfect.
But it shouldn't be random.
The goal isn't to process more candidates. It's to recognize the right ones when they appear.
---
# MiraDay Blog
Source: https://blog.mira.day/en/blog
> Product updates, recruiting perspectives, and notes from building MiraDay. Early thoughts and field reports from the MiraDay team.
MiraDay Blog [#miraday-blog]
Product updates, recruiting perspectives, and notes from building MiraDay. Early thoughts and field reports from the MiraDay team.
More posts [#more-posts]
What we're writing about [#what-were-writing-about]
---
# Recruiting Was Never One Job
Source: https://blog.mira.day/en/blog/why-a-team-of-agents
> Good recruiting is never a straight line from intake to sourcing to evaluation. The facts you uncover and the market insight you gather at each stage can send you back to revise the stage before it, so we built Mira to lead a team of agents that collaborate, debate, and challenge one another, rather than one agent that does it all.
While building MiraDay, one question we kept coming back to (and talked through with a lot of great recruiters) was how good recruiting actually happens.
When a recruiter takes on a difficult role, the work usually starts with intake. A JD on its own is rarely enough. You need to understand why the company is hiring, what the team is trying to achieve, what the hiring manager really means by each requirement, which parts of the brief are essential and which are actually negotiable, what a successful hire even looks like, and what the sourcing strategy should be.
Only then does sourcing really begin. And as it goes, the search keeps changing direction. Sometimes it even changes your original understanding of the role.
At the start of a search, you might have a very specific ideal candidate in mind. After some time in the market, you may find that the strongest candidates look quite different from that picture, or that a requirement you thought was essential is narrowing the search too far. This is especially common for hard-to-fill technical roles and senior leadership searches.
So you widen the search and start studying each candidate you find more closely. Some people whose résumés looked promising turn out to be the wrong fit once you read their actual experience; others whose résumés looked unremarkable turn out to be a much better fit than expected. And each of these discoveries changes the search again.
This is what good recruiters actually do. The process isn't a straight line from intake to sourcing to evaluation. What you discover at one stage changes how you handle the next, and sometimes it sends you back to the start to rethink the requirements.
AI recruiting often starts from a different assumption [#ai-recruiting-often-starts-from-a-different-assumption]
Most AI recruiting products today hand the entire workflow to a single agent. You describe a role; the agent interprets the requirements, searches for candidates, evaluates what it finds, and returns a list of candidates.
But that approach has a problem, and it gets more obvious as the search gets harder.
The same agent that interprets the role and decides the search strategy and direction is also the one executing the search. Then it evaluates the candidates it found against the assumptions it made at the very beginning. If that first understanding is slightly off, the whole process just follows the wrong direction, very efficiently.
A recruiter usually has ways to catch this. The market itself pushes back on the original brief. Colleagues and the hiring manager question the search strategy. Reading a candidate's résumé closely can reveal that a requirement wasn't quite right. These constant signals help a recruiter adjust, communicate, and recalibrate in time, so they end up finding the right person.
When all of that work runs inside a single agent, those moments of friction are easily lost. And we think that friction is exactly what helps you find the right person.
We looked at how recruiting teams already work [#we-looked-at-how-recruiting-teams-already-work]
This is where the idea of the Agent Team came from.
We didn't start by asking, "How do we make one agent do more?"
We asked: if we assembled a recruiting team to deliver a difficult role, how would we want that team to work?
The answer was familiar. You divide the work sensibly: someone truly understands the role before the search runs too far, someone goes into the market to see who's actually there, and someone studies candidates closely rather than assuming a good search result equals a good candidate.
The key part is that each member has their own focus and they collaborate in different ways. They don't work in isolation. The person doing sourcing might come back and say, "I'm not finding the kind of people we expected. Maybe we should rethink this requirement." The person evaluating candidates might notice a pattern that changes what the whole team should be looking for. That's how recruiting gets better, step by step.
So in MiraDay, we build agents around these different parts of the work, and keep them tightly connected as teammates.
Mira, Atlas, Scout and Vera [#mira-atlas-scout-and-vera]
**Mira** is the "brain" of the Agent Team. It leads the whole task, decides how to break it down and plan it, and judges when a search needs more than one kind of thinking.
**Atlas** handles intake. It researches the company, the market, the competition, and pay, then turns the initial request into a Role Brief: what this role actually needs (the business context, success criteria, constraints, assumptions, and open questions behind the hire). From that brief it derives an Ideal Candidate Profile: the concrete must-haves, nice-to-haves, and deal-breakers the search will run on. The point isn't to produce a more polished JD; it's to make the assumptions behind the hire explicit enough that they can actually be examined.
**Scout** takes that understanding into the market. It searches across multiple data sources for people who might fit the Ideal Candidate Profile and gathers them into a candidate pool. The facts it turns up can also push back on the profile, and sometimes on the Role Brief behind it.
If the strongest candidates keep coming from somewhere unexpected, that itself is useful information. The answer isn't always to search harder; sometimes the right move is to step back and reconsider the role.
That candidate pool isn't a shortlist. A person decides who makes the longlist, and **Vera** takes it from there. It keeps researching those candidates in depth to build the evidence a decision can rest on, prepares personalized outreach, and supports first contact and initial screening.
What Vera finds can also feed back into the search. That's by design.
These teammates aren't meant to move through a fixed sequence, each handing its output to the next. They're meant to influence each other as the search develops.
The important part isn't the number of agents [#the-important-part-isnt-the-number-of-agents]
We don't think three agents are inherently better than one.
What's actually interesting is what happens when different parts of the recruiting process are allowed to question each other.
What the market shows you reshapes the Ideal Candidate Profile. What you learn from one candidate shifts the search. And a closer look at a single person can change your whole sense of what the right candidate looks like.
That's also why we keep all the work visible in one channel. You can see what the team is finding and where its thinking shifts, instead of receiving a final list from a black box.
The goal isn't to make the process look more complicated. It's to make the reasoning behind a search more fair and transparent: something you can trust, check, and trace.
The team finds. You decide. [#the-team-finds-you-decide]
The Agent Team can take on a lot of the time-consuming work in a search. It can research the role, map the market, investigate candidates, and organize the evidence, but it doesn't make the final hiring decision.
The decision stays with the recruiter.
We think this distinction matters, because "finding information" and "judging what that information means" are two different things. AI can help you see a market far larger than you could reach alone. It can also give you more time for the parts of recruiting that genuinely need experience and judgment.
That's how we think about where MiraDay belongs: always around the tasks that help a recruiter make a better decision.
When Mira brings in the team [#when-mira-brings-in-the-team]
Most searches don't need a whole team.
Mira can handle a straightforward search on its own. When a search gets complex enough that its assumptions need testing from different angles, Mira brings in the Agent Team.
We're still working out when the Agent Team should step in, and how its members should work together. As we see more real searches and understand better where recruiters actually get stuck, the team will keep changing.
The idea behind all this exploration is simple. We don't want one agent to become a bigger and bigger version of a recruiter who does everything alone. We want Mira to work more like a good recruiting team does: understand the role first, learn from the market, look closely at the people you find, and let what you learn change what you do next.
Because finding the right person was never a straight line.
**The team finds. You decide.**
[Sign up for MiraDay](https://www.mira.day) and see how Mira runs a search.
---
# Does Talent Sourcing Still Need to Be Done by Hand?
Source: https://blog.mira.day/en/blog/why-keyword-search-fails-in-recruiting
> The problem with recruiting tools isn't that search is bad. It's that searching is the wrong job for a recruiter to be doing.
Here's something most recruiting tools get wrong: they try to make search better. Better filters. Smarter keywords. AI-enhanced Boolean strings.
But the problem was never search quality. The problem is that a recruiter is the one doing the searching.
Think about your last tough req. Maybe it was a VP of Engineering for a Series B company, and the Hiring Manager said "I'll know the right person when I see them." You spent an afternoon building Boolean strings, scanning 200 LinkedIn profiles, opening 40 tabs, and ended up with six people worth reaching out to. Three of them had already been contacted by other firms that month.
That entire afternoon wasn't recruiting. It was data retrieval dressed up as talent Sourcing.
The actual recruiting: decoding what the Hiring Manager really means by "someone who can scale the team," reading between the lines of a candidate's career moves, knowing that the person who looks perfect on paper has been quietly interviewing for three months already. None of that happened while you were scrolling through search results.
Searching is the wrong job for a recruiter [#searching-is-the-wrong-job-for-a-recruiter]
Every generation of recruiting tools has tried to improve the same workflow:
LinkedIn Recruiter made Boolean search faster. AI sourcing tools made it "smarter." But the fundamental model hasn't changed: **you are the operator, and the tool is the search engine.**
This is like giving a CEO a faster spreadsheet when what they actually need is a CFO. The bottleneck isn't the tool. It's the fact that a human is doing work that doesn't require human judgment.
Break down what actually happens during those 2-3 hours of daily Sourcing:
1. Translating a JD into search terms: **mechanical.** You're converting "entrepreneurial mindset in a regulated industry" into Boolean logic. It's lossy every time.
2. Scanning profiles to check basic fit: **mechanical.** Title, years of experience, company tier. You can assess this in seconds per profile, but multiply that by 200.
3. Reading career histories to evaluate trajectory: **partially judgment, mostly pattern matching.** Is this person on an upward arc? Do the company moves make sense? You're good at this, but 80% of it is recognizing patterns you've seen hundreds of times.
4. Compiling a candidate pool: **mechanical.** Copy, paste, format, annotate. Admin work.
Steps 1, 2, and 4 are pure execution. Step 3 is where it gets interesting, but even experienced recruiters will admit that most of the "evaluation" at the Sourcing stage is pattern recognition, not deep judgment. The deep judgment comes later, in conversation.
What if a recruiter never had to search? [#what-if-a-recruiter-never-had-to-search]
Consider a different model:
The recruiter's role shifts from **operator** to **decision-maker**. Your time goes to the work that actually requires you:
* Getting the real brief from the Hiring Manager, the one behind the JD. The one where they tell you the last person in this role was too "corporate" and they need someone who's built something from nothing.
* Evaluating cultural fit and motivation. The candidate who looks strong on paper but has jumped three times in two years. Is that a red flag, or did they have legitimate reasons? That takes a conversation, not a search result.
* Working the candidate Pipeline. The passive candidate who's happy but might move for the right story. The one who turned down a similar role six months ago but whose situation has changed.
* Telling the Hiring Manager what the market actually looks like. When the req has been open for 60 days because the comp band is 20% below market, that's a conversation only you can have.
The execution (searching, Screening, compiling) is handled by an Agent that works autonomously. Not an AI assistant you have to prompt and steer. An Agent that takes a brief, goes and does the work, and comes back with candidates.
Why this wasn't possible before [#why-this-wasnt-possible-before]
Three things changed.
**Agents can now execute multi-step tasks end to end.** Earlier AI tools could answer a question or generate a Boolean string. They couldn't independently plan a search strategy, run it across multiple data sources, evaluate each result against nuanced criteria, and deliver a structured candidate pool. That's a fundamentally different capability than autocomplete.
**Matching technology moved past keywords.** When you tell a colleague "I need someone with 0-to-1 experience at a growth-stage company," they know what you mean. They think about founders who scaled an early team, product leaders who built from scratch, operators who joined pre-Series A. A keyword search returns results containing those words. Reasoning-based matching does what your colleague does. It interprets intent, not literals.
**The economics caught up.** Two years ago, having AI reason through each candidate profile was prohibitively expensive. Inference costs have dropped 50-70% per year. Today, an Agent can evaluate thousands of candidate profiles for less than what you'd spend on coffee during a manual Sourcing session.
What changes for recruiters [#what-changes-for-recruiters]
When Sourcing is handled by an Agent, the value equation shifts.
A recruiter who spends three hours a day on Sourcing is, economically speaking, selling time. A recruiter whose Agent handles Sourcing is selling something harder to replace: market knowledge, candidate relationships, the ability to read a team dynamic and know who would actually succeed there.
Time is capped. You can't source more than the hours allow. Expertise (knowledge of which companies are in hiring freezes, which candidates are quietly looking, what compensation packages are closing deals in a specific market) is what makes a senior recruiter irreplaceable. That expertise doesn't come from search results. It comes from years of conversations, closed placements, and pattern recognition that no profile database captures.
The senior recruiters we've talked to aren't worried about this shift. One told us: "I've been waiting for this. I got into recruiting to advise clients and close candidates, not to build Boolean strings." Another put it more bluntly: "I haven't learned anything new from a search tool in five years. I've learned things from every candidate conversation."
What this looks like in practice [#what-this-looks-like-in-practice]
At MiraDay, Mira runs the Sourcing workflow end to end, bringing in an Agent Team when a search is hard:
1. You describe a role in natural language, including implicit criteria like "startup DNA," "has built a team from scratch," or "strong in regulated environments but not too corporate." The kind of brief you'd give a trusted colleague, not a search engine.
2. The Agent interprets your requirements, searches across multiple data sources, and evaluates candidates using reasoning-based matching. It understands that "0-to-1 experience" means something different from "worked at a startup."
3. You get a pool of qualified candidates with summaries, career context, and contact details, typically in a few minutes.
4. You refine through conversation: "more startup backgrounds," "include candidates open to relocation," "actually, deprioritize anyone currently at a FAANG." The Agent adjusts, not by rerunning a search, but by re-evaluating against your updated criteria.
During internal testing, recruiters running real Sourcing tasks have reported significant time savings on candidate research.
If you're spending hours every day on manual Sourcing (hours that could go to Hiring Manager calls, candidate conversations, and closing), that time is recoverable.
[Sign up for MiraDay](https://www.mira.day) and try Agent-powered Sourcing for yourself.