Hire developers who actually know the code.
Attest reads a candidate's real repositories, checks they understand their own code, and hands you a ranked shortlist, with the file and line behind every skill. Free for candidates, and free for recruiters during our pilot.
Illustrative. A real score comes from a candidate's own repositories, and thin evidence caps the result.
Verified from the places developers already build
See how scoring works- GitHub
- LeetCode
- CodeChef
- HackerRank
- Naukri
GitHub and LeetCode are scored. CodeChef and HackerRank are shown. The Chrome extension adds verification to LinkedIn and Naukri profiles.
Attest turns a developer's real code into proof a recruiter can check.
It reads a candidate's repositories, asks them to explain their own code, and our proctoring software prevents cheating. Recruiters get a ranked shortlist, with the file and line behind every skill.
of a candidate's real work are read in depth: structure, tests and who wrote the commits.
are detected from real imports and API usage, never from a bio.
points, split into six parts a recruiter can audit one by one.
Every skill opens the exact file and line in the candidate's own code.
From a GitHub link to a shortlist.
Four steps, and every one of them leaves evidence a person can check.
Connect GitHub and LeetCode once. Attest reads your real work, shows exactly why you scored what you did, and keeps your profile current as you ship.
- Expressshop-api/src/server.ts · line 12
- PostgreSQLshop-api/src/db.ts · line 4
- Redisshop-api/src/cache.ts · line 9
- Dockershop-api/Dockerfile · line 1
Every skill keeps the file and line it came from, so a recruiter can open it and check.
In db.ts you open a pool once and reuse it. What would go wrong if each request created its own connection?
Questions about your own code, with our proctoring software preventing cheating. Odd patterns are flagged for a person to review, never auto-rejected.
Describe the developer you need in plain English. Attest ranks candidates by proven skills, explains every match, and gives a non-technical recruiter a credible way to screen.
Every result says what is proven, what is only related, and what has no evidence yet.
- Shortlisted100%
- Interviewed62%
- Offered28%
- Hired18%
Sample data, for illustration.
Shortlist, message and pitch in one place. After 20 recorded outcomes, ranking adjusts a little, and explains how.
Students join with consent after seeing exactly what is shared, and can leave any time. You see who has verified, who has not started, and which skills the batch is missing.
Batch dashboard
Verification rates, code-ownership results, common skills and the biggest gaps for a role you choose. Export everything as CSV.
Verified badge
Students can add an opt-in badge to their README that links to a public profile. It never includes contact details, notes or answers.
[](https://getattest.dev/v/<code>)
Browser extension
A Chrome extension shows a candidate's public verification right on LinkedIn and Naukri profiles, so recruiters see the evidence where they already work.
Built for the people on both sides of the hire.
Connect once and your score updates itself.
For every candidate.
See why you scored what you did, and what to build next. Let your work speak.
Export and delete everything in Settings.
An opt-in badge that never shows your contact details.
Free for candidates, and free for recruiters during our pilot. Colleges and agencies, talk to us.
Every skill links to a line of code.
Skills come from imports and real API usage, not from a bio. Open one and you land on the exact file in the candidate's repository.
1 import express from 'express';2 import { Pool } from 'pg';34 app.get('/api/orders', async (req, res) => {5 const rows = await pool.query(SELECT_ORDERS);6 res.json(rows);7 });
- Expressline 1
- PostgreSQLline 2
- REST API designline 4
- SQLline 5
- Plain-English search
- Describe the person you need. Results say why each one matched.
- Interview kits
- Questions from the candidate's own code, with what a good answer sounds like.
- Private notes
- Your notes help your team and never lower a candidate's score.
- Whole batches
- Colleges and clubs onboard a cohort with one invite link.
A score you can audit.
Candidates see the same breakdown recruiters do, with the reasoning for every point. Copied code, tutorial clones and starred empty repos score low, and thin evidence caps the result.
Competitive programming counts too, but only from a verified LeetCode account.
- Project substance250
- Engineering practice200
- Code quality200
- Sustained activity150
- Breadth and depth150
- Community impact50
Rookie · Rising · Advanced · Elite · Legendary
Every score comes with its evidence.
Six checks sit behind each score. A recruiter can open and inspect every one of them.
Scores don't measure teamwork. A score is advice for a person, never the decision.
- OAuth
The account is theirs
GitHub sign-in confirms they own the account. LeetCode is verified with a one-time token. One account backs one profile.
- Fixed rules
Skills come from code, not words
About 60 skills are detected from real imports and API usage by fixed rules. Each keeps its file and line.
- Commit history
We check who wrote it
Commit authorship is read, so forks, tutorial clones and copied code score low. Stars and followers barely count.
- Proctored
They must understand it
Questions about their own code, graded against that code, with our proctoring software preventing cheating.
- Human review
Flags go to people
Typing patterns and an AI-writing estimate flag answers for review. A flag starts a conversation and never rejects anyone.
- Evidence cap
Thin evidence caps the score
A few small repositories can't produce a high score, and the profile says when evidence is thin.
Hiring data is about real people. We act like it.
- You stay in control
- Hide your profile any time, keep the public badge off, and download or delete everything from Settings.
- Scores can't be edited
- Only our servers write scores and verification flags. Login tokens never reach the browser.
- Flags are signals, not verdicts
- AI-writing estimates are unreliable, so they only ask a human to look closer.
Short code excerpts are reviewed by Google Gemini. We keep derived skills and scores, not copies of your repositories. Privacy policy
Questions, answered plainly.
They don't have to trust us. Every skill opens the exact file and line in the candidate's repository, and the breakdown shows the reasoning for every point.
They sign in with GitHub OAuth, so GitHub itself confirms it. LeetCode is verified with a one-time token in the bio. An account links to one profile only.
Our proctoring software is built to prevent cheating, and it flags suspicious answers for review along with an AI-writing estimate. No tool can prove AI use, so a flag is a reason to talk, never an automatic rejection.
A thin profile gets a low score, and the profile says evidence is thin. Every candidate sees what to build next, and the walkthrough result is shown separately, so understanding counts too.
Candidates can give read-only access to the private repositories they choose. Those are read the same way, and only derived skills and scores are kept, never a copy of the code.
GitHub and LeetCode are scored today. CodeChef and HackerRank are shown but not scored. GitLab and Bitbucket are not supported yet.
Free for candidates, and free for recruiters during our pilot. Colleges and clubs can talk to us about batches.
No. Scores are evidence to help a person decide faster and more fairly. The decision stays with people.