AI-assisted recruiter screening

Turn technical screening calls into clearer client submissions.

Run the screen in your browser or paste the transcript afterwards. Veyrag keeps what the candidate actually said, separates it from what they didn't, and you approve what goes to the client.

  • 10 free recruiter screens with a confirmed email address
  • No credit card required
  • Never joins your Zoom, Teams or Meet call

The walkthrough was made in October 2026 from illustrated screens with fictional data, so it looks different from today's app.

What Veyrag does

Turn recruiter screening conversations into accurate, recruiter-approved client submissions.

A résumé says one thing.

The screen hears another.

The client gets a few notes.

Veyrag sits beside the recruiter. It prepares the screen from the role, keeps what the candidate actually said, and turns it into a submission you approve, so the client gets what the screen really showed.

A recruiter reading a candidate's résumé across the table, in black and white.

How it works

Prepare, screen, review the evidence, approve and share.

You lead every step. Veyrag never joins the meeting and never sends anything to a client.

  1. 01

    Prepare

    Add the job description and the résumé. Veyrag maps the role's requirements and plans what a 10, 15 or 20-minute screen can cover.

  2. 02

    Screen

    Open Live Screen in your browser beside the call: your microphone and the meeting tab's audio, in Chrome, with a useful next question suggested when one is ready. Or paste the Zoom, Teams or Meet transcript after the call. Veyrag never joins the meeting.

  3. 03

    Review evidence

    A Candidate Brief built on the candidate's own words: what they described doing first-hand, what they described as team work, what they said they haven't done, and what wasn't tested. Mark anything overstated or in the wrong area.

  4. 04

    Approve and share

    Client-ready wording grounded in the screen. You write the recommendation, approve a version, then copy it into your ATS or share a link you can revoke. Veyrag never sends anything to a client.

The Veyrag companion during a live screen: a suggested question with Why this, and Ask this, Not now and Other topic.
The live companion during a recorded screen, with a suggested question. Fictional candidate.

A complete example

One screen, from résumé to approved submission.

Jordan Vale is a fictional Senior Backend Engineer. Everything below comes from one test screen run through Veyrag: the résumé and job description written for the test, what Jordan said on the call, Veyrag's evidence record, and the submission a fictional recruiter approved. The text is quoted unedited; only the layout is ours. Nothing was registered or signed in to show it.

  1. 1 · Before the screen

    The résumé and the job description

    Résumé (fictional)

    “Helped add consumer-lag monitoring in Grafana and on-call alerts to surface delayed event processing.”
    “Operations: Grafana, Prometheus, Docker, GitHub Actions, structured logging”

    Job description · helpful, but not required

    “Kubernetes, AWS, Grafana, OpenTelemetry, CI/CD pipelines and familiarity with system design reviews.”

    The role asks about Kubernetes. Nothing in the résumé mentions it.

  2. 2 · In the screen

    What Jordan said

    RecruiterHow did you know when the consumers were falling behind?

    Jordan“We watched consumer lag in Grafana. I helped configure an alert so the on-call engineer got notified if the lag stayed above our threshold for about five minutes…”

    RecruiterHave you directly mentored junior engineers in a formal or ongoing capacity?

    Jordan“Now, I haven't mentored engineers directly. I've answered teammates' questions, but I haven't owned a mentoring program…”

    Kubernetes never came up in the conversation. The live transcript heard “Now” for “No”.

  3. 3 · Veyrag's evidence record

    What the screen established, and what it didn't

    Veyrag · Recruiter Screen Intelligence · internal, recruiter only

    Role requirementReadingFrom the record
    Kafka event streamingFirst-hand account“I implemented a consumer in our order events service. It read order status events from a Kafka…” Q2
    Cloud observability deliveryNot established Collaborative accountGrafana came up within team work; the requirement was not assessed. Not discussed: Kubernetes, AWS, OpenTelemetry, CI/CD. Q3
    Technical leadership mentorshipExplicit boundaryStated boundary: direct mentorship of less-experienced engineers. Q7

    Recruiter-only · screen-level evidence · not an independent skills verification or a hiring recommendation

  4. 4 · Recruiter-approved

    The client submission, version 3

    Approved Approved by the recruiter on Nov 13, 2026 (version 3). Excerpt.

    Jordan Vale · Senior Backend Engineer

    Highlights from our screen

    Production incident troubleshooting. They traced repeated event IDs across the consumer logs.

    Production Kafka. They implemented a consumer in the order events service. It read order status events from a Kafka topic, checked the event ID to avoid processing duplicates, and wrote the result to a downstream status service.

    Recruiter recommendation

    Recommend a technical interview. Jordan gave first-hand accounts of Kafka consumer work, Java and Spring Boot service changes and production incident troubleshooting. Please explore PostgreSQL ownership, service-resilience ownership and REST API authentication. Kubernetes and cloud observability were not discussed in this screen.

    Highlights summarize the candidate's own account in a recorded screen. They are not an independent verification of skills, identity or employment.

What the client sees. The role asked about Kubernetes; the résumé lists Grafana. On the call Jordan described Grafana alerting as team work and never mentioned Kubernetes. The record says exactly that, the mentoring answer stays the candidate's own boundary rather than a weakness, and the approved submission credits only what Jordan described.

The Candidate Brief

Every conclusion sits next to the candidate's own words.

What the screen established, what the candidate said they haven't done, and what wasn't tested are kept apart. Evidence states, never scores.

  • EstablishedA concrete, first-hand account with a detail the next interviewer can check.
  • Some evidenceRelevant, with an important part still to validate.
  • Mentioned onlyNamed, but not explored.
  • Said noThe candidate said they haven't done it: a stated boundary, not a weakness.
  • Not testedThe screen didn't reach it. No conclusion either way.

Was Veyrag right? Correct, Overstated, Wrong area or Needs context. Your review stays with the record; the evidence itself is never rewritten.

The evidence summary of a Candidate Brief from a recorded live screen: Kafka established with the candidate's quote, areas with some evidence, and two areas the candidate said they have not done.
A Candidate Brief from a recorded live screen. Fictional candidate.
An approved client submission, version 6, with the exact approved text: a summary and highlights in the candidate's words.
An approved client submission. Fictional candidate.

The client submission

Client-ready, and recruiter-approved.

Each strength rests on something the candidate actually said. Veyrag suggests the wording; you choose it, write the recommendation yourself, and approve a version.

  • Needs recruiter review → Draft → Approved. Always clear where it stands.
  • The exact approved text. Copy it into your ATS, download it, or share a link you can revoke.
  • Nothing sent for you. Veyrag never sends anything to a client. Every approved version is kept, and you can withdraw one.

Audio recording import is in testing and is not yet available. Technical Interview is coming soon. Veyrag organizes evidence from the conversation; it does not verify a candidate's abilities or make hiring decisions.

A recruiter interviewing a candidate in an office, seen through glass, in black and white.

Why staffing agencies use Veyrag

Your clients get what the screen actually showed.

  • Submissions you can stand behind

    Every strength is the candidate's own words, taken from the conversation itself.

  • Honest about the gaps

    What wasn't reached is listed as not tested, and what the candidate said they haven't done is their own statement. Never a weakness, never a guess.

  • The recruiter stays in control

    You lead the call, correct the readings and approve every version. Veyrag never joins the meeting.

  • Your ATS stays the system of record

    Copy the approved text where it belongs. Veyrag isn't trying to replace it.

FAQ

Questions recruiters ask first.

Something else? Email support@veyrag.com.

A recruiter interviewing a candidate in a modern office, in black and white.
What is Veyrag?

A recruiter-screening copilot. It prepares a recruiter screen from the résumé and job description, supports the screen live or imports its transcript, organizes your notes and the candidate's essentials, and drafts a client submission grounded in what the candidate said, for you to review and approve.

Who is it for?

Recruiters and staffing agencies who screen candidates for technical roles and submit them to clients or hiring managers. During early access, accounts are for agencies operating in the United States, screening candidates based in the United States, signing up with an email address they confirm, work or personal.

What meeting audio is supported?

For a live screen, Veyrag uses your microphone for your side and the audio of the browser tab your meeting runs in (Zoom, Teams or Meet in the browser) for the candidate's side, in Chrome on a desktop. No bot joins the call. If the meeting runs in a desktop app, import its transcript afterwards instead.

Can I import an existing transcript?

Yes. Upload a transcript export (.vtt, .srt or .txt) or paste it, with each line starting with who spoke. You confirm which speaker is the recruiter and which is the candidate, correct any line, record the call's date, and finish the import. One import uses one free screen.

Are audio recordings supported?

Not yet. Importing an audio recording is in testing and is not available in the product today. Until it is, import the recording's transcript.

Does Veyrag verify skills?

No. Veyrag organizes what the screen established at screen level: what the candidate described, what they said they haven't done, and what wasn't tested. It does not verify technical ability, score candidates or make hiring decisions. You review every reading and approve anything that goes to a client.

How many free screens are included?

Each eligible account gets 10 free recruiter screens once, for life, after its email address is confirmed: no credit card, no expiry. A free screen is one live Recruiter Screen of 10, 15 or 20 minutes, or one imported transcript.

Temporary or disposable email addresses can be used to register, but they do not receive automatic free credits. You can request a manual review for free screening credits.

What happens to candidate data?

Screening records stay under your account and are deleted 30 days after the candidate's recorded consent, or earlier when you delete them. Veyrag does not keep call audio; audio and text are processed by OpenAI for transcription and analysis under its API data policy. The details are in the Privacy Notice and the Data Processing Terms.

Start with 10 free recruiter screens.

No credit card required. Import a screen you already ran, or run your next one with Veyrag.