Browser screen
First-pass filtering with the evidence timeline above. Nothing to install, and candidates see exactly what's logged before they start.
Start free →Built by an engineer who's run 400+ technical interviews
A technical screen that actually screens for the job.
Real code. Real test cases. Sandboxed dev servers in the browser — wired to public job postings so candidates go from apply to screened in one sitting. And your whole team reviews live, in one room.
- ✓ Free up to 5 candidates / month
- ✓ No credit card, no demo call
- ✓ No proctoring spyware
How it works today
One link does the whole top of funnel. Apply, screen, rank.
Post a role to your Kendor board. The instant a candidate applies, their coding screen is sent automatically — if you've enabled it. No ATS to wire up, no scheduling, no manual hand-off.
ASacme.systems · Senior FrontendLive · accepting applicants
Step 1 of 5
Public job post
A candidate finds your open role on your own Kendor careers board — one public link, your handle.
Applicant pipeline
6Applied
0Screening
1Reviewed
1Shortlist
Auto-advancing — tap any step to explore. Auto-screen is a per-role toggle; leave it off and applicants simply queue for review.
Two sides, one pipeline
Built for the people on both ends of the screen.
For candidates
Apply once, screen immediately, and never install a thing.
- 1Apply from the job postOne link, no account hoops. Reuse your profile and résumé.
- 2Solve it in the browserA real sandboxed dev server. Write code, run the tests, see them pass.
- 3Walk it through liveInstead of a repeat interview, join the review room and defend your own code.
No webcam. No screen recording. No spyware install on the web screen.
For hiring teams
Engineers grade, recruiters route — one review pipeline for both.
- 1Post the role, set the screenAttach a challenge once. Toggle auto-send on apply per role.
- 2Review, don't re-runReplay how they got there: run history, test results, focus events.
- 3Decide togetherPull the whole panel into one live room. Follow each other's cursors, huddle, keep notes staff-only.
Per-reviewer pricing — you're never billed per candidate.
New in the workspace · chat
Live todayEvery candidate is one thread.
Messages, screens, interviews — the whole pipeline lands in a single conversation per candidate. No shared mailbox, no reconstructing what happened from four pages.
'%3e%3cpath%20d='M50%200H14C6.26801%200%200%206.26801%200%2014V50C0%2057.732%206.26801%2064%2014%2064H50C57.732%2064%2064%2057.732%2064%2050V14C64%206.26801%2057.732%200%2050%200Z'%20fill='%230E1626'/%3e%3cpath%20d='M22%2019.5C22%2018.1193%2020.8807%2017%2019.5%2017C18.1193%2017%2017%2018.1193%2017%2019.5V44.5C17%2045.8807%2018.1193%2047%2019.5%2047C20.8807%2047%2022%2045.8807%2022%2044.5V19.5Z'%20fill='url(%23paint0_linear_54914_121)'/%3e%3cpath%20d='M29.664%2045V19.44H35.136V30.6L44.424%2019.44H50.796L41.4%2030.708L51.372%2045H45.144L37.8%2034.38L35.136%2037.476V45H29.664Z'%20fill='%23FBF9F4'/%3e%3c/g%3e%3cdefs%3e%3clinearGradient%20id='paint0_linear_54914_121'%20x1='17'%20y1='17'%20x2='17'%20y2='47'%20gradientUnits='userSpaceOnUse'%3e%3cstop%20offset='0.5199'%20stop-color='%233856DC'/%3e%3cstop%20offset='0.52'%20stop-color='%23D24327'/%3e%3cstop%20offset='1'%20stop-color='%23D24327'/%3e%3c/linearGradient%3e%3cclipPath%20id='clip0_54914_121'%3e%3crect%20width='64'%20height='64'%20fill='white'/%3e%3c/clipPath%3e%3c/defs%3e%3c/svg%3e){kind=small}
Inbox / Ada OkaforSenior Backend · ShortlistSubmission received — 5 of 5 tests · Jul 14 View playback
Nice run, Ada. We'd love you to walk it through live — does Thursday 2pm work?
MCAO
Works for me. Anything I should prepare?
Live review — Thu 14:00
Fresh link, expires after the sessionJoin opens Thu
Message Ada — she gets an email tooInvite to live review
Interview mode · live review
In private betaReview together. Admit, follow, pair.
The same pipeline, live — the whole assessment, every challenge, in one room. Reviewers and the candidate share the same bench: presence, live cursors, comments you can answer out loud, and a huddle one click away.
'%3e%3cpath%20d='M50%200H14C6.26801%200%200%206.26801%200%2014V50C0%2057.732%206.26801%2064%2014%2064H50C57.732%2064%2064%2057.732%2064%2050V14C64%206.26801%2057.732%200%2050%200Z'%20fill='%23FBF9F4'/%3e%3cpath%20d='M22%2019.5C22%2018.1193%2020.8807%2017%2019.5%2017C18.1193%2017%2017%2018.1193%2017%2019.5V44.5C17%2045.8807%2018.1193%2047%2019.5%2047C20.8807%2047%2022%2045.8807%2022%2044.5V19.5Z'%20fill='url(%23paint0_linear_54914_117)'/%3e%3cpath%20d='M29.664%2045V19.44H35.136V30.6L44.424%2019.44H50.796L41.4%2030.708L51.372%2045H45.144L37.8%2034.38L35.136%2037.476V45H29.664Z'%20fill='%230E1626'/%3e%3c/g%3e%3cdefs%3e%3clinearGradient%20id='paint0_linear_54914_117'%20x1='17'%20y1='17'%20x2='17'%20y2='47'%20gradientUnits='userSpaceOnUse'%3e%3cstop%20offset='0.5199'%20stop-color='%233856DC'/%3e%3cstop%20offset='0.52'%20stop-color='%23D24327'/%3e%3cstop%20offset='1'%20stop-color='%23D24327'/%3e%3c/linearGradient%3e%3cclipPath%20id='clip0_54914_117'%3e%3crect%20width='64'%20height='64'%20fill='white'/%3e%3c/clipPath%3e%3c/defs%3e%3c/svg%3e){kind=small}
Review / Senior Backend / Ada OkaforLiveMCDPStart huddlerate-limiter / src / allow.ts · read-only
import { redis } from './redis';
export async function allow(key: string) {
const count = await redis.incr(key);Devin
return count <= LIMIT;
}
MCWhat happens here when Redis is down? · line 4 · reply live or async
AOAda Okafor is asking to joinWaiting in the green room · devices checked
AdmitNot yet
Ada is knocking — waiting in the green roomroom review:asmt_8Hq2 · 3 challenges
One room per assessment — every challengeReview is read-only — cursors are decorationsStaff-only notes on every surfaceCandidates never see scores, rubrics or other candidatesAgents are org-provided, enabled per assessmentRequest beta access →
Coming next · the agent tab
Coming soonAn agent in the bench. On your terms.
A chat tab beside the editor where candidates drive Claude Code or Codex against the real workspace — the way they'd work on the job. Your org provides the credentials and switches it on per assessment.
- 1Org-provided, org-enabledYour credentials, your call, per assessment. No agent unless you turn it on — and Kendor never supplies the AI itself.
- 2Private while they solveThe chat is the candidate's own space during the session — nobody shoulder-surfs their prompts live.
- 3Every prompt in playbackThe full transcript lands in the reviewer timeline. What they asked, what they accepted, what they rewrote — that's the signal.
We don't block AI, and we don't smuggle it in. If working with an agent is part of the job, screen for it — if it isn't, leave the tab off.
Integrity
Evidence, not verdicts.
We don't compute an “integrity score” and we never auto-reject. The reviewer sees what happened, timestamped, and decides what it means. Facts on a timeline — the conclusion stays yours.
On every web screen today
- ✓Focus events — every tab-leave and fullscreen exit, logged with the submission.
- ✓Paste capture — external pastes stored verbatim for the reviewer.
- ✓Full playback — every keystroke, run and marker replayed on one scrubber.
Landing next
Coming soon- ○Away durations — “left the tab 7 times” becomes “away 4m 32s at 00:17”.
- ○Active vs elapsed — “38m active of 55m”, computed from the playback stream.
- ○Paste survival — did that paste ship verbatim, get adapted, or get thrown away?
- ○Second-device facts — noted server-side when one invite is active from two clients at once. Hashed identifiers, never raw IPs.
- ○Similarity facts — fingerprint overlap with other submissions on the same challenge, starter code excluded. Shown next to the grade, never in it.
Six things we won't build. Ever.
Surveillance theater punishes honest candidates and doesn't stop a determined cheat. When the stakes justify more, use a tier built for it — below.
- ×Webcam and ID checks
- ×Screen recording
- ×Composite “integrity scores”
- ×Auto-reject rules
- ×Copy-block and devtools theater
- ×Spyware installs
For take-homes
A menu-bar companion that records the process — pasted sources, searches, AI usage. Reaching for Claude counts in a candidate's favor. No camera, no mic.
Notify me →For final rounds
Locked full-screen for when a bad hire is expensive. Real OS runtimes — Node, Docker, PHP — allowlisted to what the assessment needs.
Request early access →Data we keep
Code, run logs, focus events, messages, reviewer comments. Deleted automatically 90 days after the role closes — or any time you ask.
Data we don't
No webcam, no audio, no screen recording on the web screen. No third-party ATS export by default.
AI usage
Your call, not ours. Orgs provide and enable agents per assessment — we don't block AI, and we don't smuggle it in. Every prompt lands in playback.
Where it runs
EU and US regions. Single tenant on Enterprise. SOC 2 in progress (Q3).
An honest comparison
vs HackerRank, CodeSignal & take-homes.
We're not for everyone. If you screen thousands of new-grads on algorithm puzzles, the big platforms are built for that. Kendor is for teams hiring on real, job-shaped work.
KendorHackerRankTake-home
What you screenReal, job-shaped codeAlgorithm puzzlesOpen project
Candidate setupNothing — browserAccount + IDETheir own machine
Apply → screenAutomatic on applyManual inviteManual email
Live reviewIn the room, togetherSeparate call toolNone
Pricing modelPer reviewerPer candidateFree (their time)
ProctoringTimestamped factsWebcam + lockdownNone
PlagiarismSimilarity facts, no verdictsAuto-flag scoresUnknown
AI usageOrg-enabled agentsBlockedUnknown
No leaderboards, no per-candidate metering — and similarity is a fact beside the grade, never an auto-fail score. If you want verdicts handed to you, we're genuinely not your tool. Read the full, honest comparison →
Security and compliance
Ready for your security review.
Candidate data is personal data. We built for EU procurement from day one — not as an enterprise upsell.
GDPR by default
A signed DPA on every plan, including free. Candidates can request an export or erasure of their data directly — no support ticket, no gatekeeping.
EU data residency
Pick the EU region and everything — code, playback, chat — is stored and processed in Frankfurt. It never leaves the EU, backups included.
Encrypted everywhere
TLS 1.3 in transit, AES-256 at rest — playback streams and chat included. Single tenant available on Enterprise.
Retention you control
Candidate data deletes automatically 90 days after a role closes — set it shorter if you like. Deletion means deletion: backups purge within 30 days.
Subprocessors, listed
A short public list with 30 days' notice before any change. No ad-tech, no analytics on candidate surfaces, and your data is never sold or used to train models.
Access controls
SSO with SAML on the Team plan, role-based access per job, and an audit log of every reviewer action — who watched which playback, when.
GDPR compliant
Regulation (EU) 2016/679
EU data residency
Frankfurt, eu-central-1
DPA on every plan
Signed with SCCs, free included
SOC 2 Type II
In progress — report Q3 2026
Security questionnaire? We fill those in fast.
Pricing
Per reviewer, never per candidate.
Honest tiers. No “contact us” gate for small teams. The Team plan covers everything for 95% of customers.
Free
Freeforever, even later
For trying it on a real role.
- 5 candidates / month
- Web screens
- Job postings + 1 board
- Community challenge library
- 1 reviewer seat
Team
Freein betathen $49 / reviewer / mo
For hiring teams running regular screens.
- Unlimited candidates
- Job postings + applicant pipeline
- Auto-screen on apply
- Interview mode + candidate chat
- Observer + desktop kiosk (on launch)
- Side-by-side comparison
Enterprise
Freein betathen custom pricing
For compliance, SSO, dedicated infra.
- SSO / SAML
- Audit logs
- Single-tenant deploy
- EU data residency
- Priority support
- Custom integrations
A note from the maker
Why I built this
I ran the technical interviews at my last 2 jobs. Around 400 of them, give or take. The same pattern every time: the take-home was either too easy and told us nothing, or too hard and ate the candidate's weekend for free. The vendor tools were worse — they made our candidates install spyware to prove they weren't cheating on FizzBuzz.
So I built the thing I wanted. Real code, no proctoring drama, runs in any browser. Then came the part I missed most from good in-person loops: sitting down with the candidate and their code, together. That's interview mode — the debrief and the walkthrough in one room, instead of another scheduled interview. It's in private beta now.
Kendor is small. It's me and a handful of engineering teams using it on real hires. If you try it and something is broken or stupid, my email is at the bottom of this page.
— Thet AungFounder · Europe
Early access
Be first to the next thing.
Job postings, candidate chat and interview mode are live today. Tell us what you're hiring for and we'll get you set up — plus first access to the agent tab, Observer and the desktop kiosk the moment they ship.
- ✓ Usually a reply within a day — from the person who built it
- ✓ No demo call required to start
- ✓ Or just email hello@kendor.io