AI Resume Screening
Paste a CV and the role you are hiring for. An AI agent scores it against a fixed rubric — the same one for every candidate — and returns a scorecard where each score cites the line it came from, so you can disagree with it specifically rather than in general.
Where a scorecard actually helps
Screening goes wrong in predictable ways: too many CVs, too little time, and a first impression that hardens before anyone reads the second page. These are the moments a fixed rubric changes the outcome.
A pile of applications
Everyone is read against the same rubric, including the ones that arrive on a Friday afternoon when your attention is gone.
A role outside your expertise
Hiring your first designer, or your first accountant. The rubric gives you a structure for judging work you cannot judge by instinct.
Career changers and gaps
Judged on demonstrated capability, not continuity. Gaps are not scored as a defect, because a gap is not evidence of anything.
Engineers with public work
Optional enrichment from a public GitHub profile, so what someone has actually built counts alongside how they described it.
Preparing the interview
The most useful output is often the list of things the CV leaves ambiguous — which is exactly what the interview is for.
Defending a decision later
An evidence-bound scorecard is a record of why, which matters when a rejected candidate, or your own team, asks.
How it works, in three steps
- 1
Give it the CV and the role
Plain text or a link to a PDF, plus the role you are hiring for. A GitHub handle is optional and adds public work.
- 2
The agent scores against a fixed rubric
The same rubric for every candidate, with each score bound to the evidence in the CV — and gaps left as gaps, not guesses.
- 3
You read it and decide
The scorecard lands in your workspace as a report, with a card on the Humans board. The decision stays yours.
Why this beats keyword filtering
Most screening automation matches keywords and rejects the rest. That is fast, and it removes exactly the candidates who describe the same experience in different words.
| Keyword filtering | Veii agent | |
|---|---|---|
| How a CV is judged | Presence of the right words | Demonstrated capability against a stated rubric |
| Why this score | Opaque — a rank with no reason | Each score cites the line it came from |
| Career gaps | Often penalised implicitly | Not scored as a defect; explicitly out of the rubric |
| Unusual backgrounds | Filtered out for not matching the pattern | Judged on what they can do, not on pedigree |
| Who decides | The filter rejects before a human reads | Advisory only — every CV still reaches a person |
| What you can show later | A ranking you cannot explain | A scorecard with the evidence for every judgement |
Who this is for
Founders hiring their first few people
Your first hires matter most and you have the least process. A fixed rubric will not tell you who to hire, but it stops the tenth CV being read more harshly than the first because you are tired.
We are a five-person company hiring our first Operations Manager. Evaluate this CV against that role, cite the evidence for each score, and tell me what to probe at interview because the CV does not answer it.
Small teams without a recruiter
Nobody here screens CVs for a living, and the role is often outside your own expertise. The value is not the score — it is having a structure that is the same for the candidate you liked immediately and the one you almost skipped.
Evaluate this CV for a Senior Backend Engineer role and enrich it from their public GitHub. I am not an engineer, so tell me what their public work shows and where it contradicts what the CV claims.
Hiring managers who have to justify a decision
Someone will ask why this candidate and not that one — a rejected applicant, a colleague, sometimes a regulator. A scorecard bound to evidence is an answer; a gut feeling recalled three weeks later is not.
Evaluate these two candidates for the same Product Designer role against the same rubric, and show me where they actually differ on evidence rather than on impression.
Frequently asked questions
What is AI resume screening?
It is using an AI agent to read a CV against a stated rubric and produce a scorecard: a score per dimension, each bound to the evidence in the CV that produced it. Done properly it is a reading aid that makes your criteria explicit and applies them consistently — not a machine that decides who gets rejected.
Is the resume screening tool free?
Yes. Veii runs on free-tier models, so evaluating a CV costs nothing, there is no per-candidate charge and no credit card at signup. That matters most for the small teams who screen a few dozen CVs a year and would never buy a hiring platform.
Does it reject candidates automatically?
No. The output is a scorecard for a person to read, and every CV still reaches a human. Automated rejection is exactly the failure mode that makes screening tools harmful — it hides the decision, removes the appeal, and makes a model’s blind spot into a policy. The decision stays with you.
How does it avoid bias?
The rubric is fixed and applied identically to every candidate, protected characteristics are ignored, career gaps are not scored as a defect, and pedigree — which university, which brand-name employer — does not earn points on its own. No screening process is bias-free, but a stated rubric with cited evidence is at least one you can inspect and argue with.
What does the scorecard actually contain?
A score for each dimension of the rubric, the specific evidence from the CV behind each one, an explicit list of what the CV does not tell you, and the questions worth asking at interview. The last two are often more useful than the scores themselves.
Can it read a PDF CV?
Yes, from a link to the PDF, which is extracted to text automatically. Scanned or image-only PDFs with no text layer are the exception and will not work — send those as text instead of getting a confident evaluation of an empty document.
Can it look at a candidate’s public work?
Optionally, yes: give it a GitHub handle and it will enrich the evaluation with public repositories. This is most useful when it disagrees with the CV in either direction — someone who undersold themselves, or someone whose described experience is not visible in the work.
Screen your next CV
Create a free workspace and run the CV you are unsure about. Read the evidence behind each score, then make the call yourself.
Evaluate a CV free