Launching today
Kendor
AI-allowed coding interviews in a real dev environment
4 followers
AI-allowed coding interviews in a real dev environment
4 followers
Coding interviews where candidates can use AI, and you see exactly how they did it. • Start from your own repo: paste a job description and Claude drafts the task • A real project sandbox: editor, terminal, database, live preview • Candidates use their own agent: Claude Code, Codex, Opencode, Cursor, Copilot and more • Replay every prompt, edit and test run, plus an AI summary • Run their code and tests yourself Screens, live interviews and hackathons. EU hosted.
Free Options
Launch Team / Built With
Render Sandboxes: a computer for every agentRun untrusted code safely. Idle sandboxes cost nothing
Promoted
Hi Product Hunt 👋 I’m Thet, and I built Kendor.
Kendor is the first technical interview platform where candidates can use their own AI tools in a secure, controlled VS Code environment online. We support Claude Code, Codex, Antigravity, Cursor, Copilot, OpenCode and Cline.
It’s also the first technical hiring platform that runs a whole hackathon inside an online IDE. Every builder codes in their own sandbox, and everyone meets in a virtual venue built on WorkAdventure, where they can walk around and talk to each other.
Hiring teams have two options: add their own API key and allow specific AI tools, or let candidates connect their own AI accounts and code with the tools they already know.
We support screens, live coding interviews with a built-in video call and whiteboard, and hackathons. Each one gives the hiring team a different way to assess candidates and different insights.
Kendor records the AI transcripts, so hiring teams can see how candidates talk to AI and how they prompt. In the era of AI-assisted coding, we believe this tells you more than their final answers alone.
We believe the best way to see how someone actually works is to give them a real coding environment, a real on-the-job problem and their own AI agent. It shows far more than algorithm puzzles, AI interviews or banning AI.
To prove it, we ran a hackathon last weekend(Kendor Build Day) with 50 builders and 15 experienced judges. Builders were dropped into real-life scenarios: fixing problems customers had reported, or adding features a client asked for, while Kendor recorded everything.
At the end, Kendor wrote an AI summary to give the judges a quick picture of what each builder did. Judges could then read the code, look through the terminal and AI transcripts, and run the tests. In the future, they’ll also be able to talk with the builder.
We’ve already had a lot of feedback from builders, in LinkedIn posts and in messages to us that they enjoyed this kind of challenge, and the Kendor platform and its IDE.
I’d love to hear from you: would you let candidates use AI in your interviews? And if you already do, how do you tell good AI use from bad?