Which coding interview platforms let candidates use AI?
CoderPad, HackerRank, CodeSignal, Codility and Kendor all let candidates use AI now. They differ in what kind of AI it is, whose account it runs on, who turns it on and what reviewers see.
Most coding interview platforms now let candidates use AI: CoderPad, HackerRank, CodeSignal, Codility and Kendor all do. The difference is what kind of AI. Most offer a chat assistant built into their editor. Kendor runs real coding agents, Claude Code and Codex, in the terminal of a full development environment, on the company's account or the candidate's own. Pick based on how your engineers actually work.
This comparison reflects each vendor's own documentation as of November 2026. Features change quickly, so check the sources at the end before you decide.
Four questions to ask any platform
What kind of AI is it?
There are three kinds, and they test different skills:
- A chat assistant next to the editor. The candidate asks questions and copies or applies answers.
- Inline completions, like Copilot in an editor.
- A coding agent that reads the project, edits files and runs commands on its own, like Claude Code or Codex.
If your team works with agents every day, an interview with only a chat box tests a different job.
Whose model and account does it run on?
Some platforms pick the model for you. Others let you bring your own provider account, or let the candidate use the subscription they already work with. That decides cost, data flow and how familiar the tool feels to the candidate.
Who turns it on, and is it on by default?
You want AI off for some interviews and on for others. Check whether you can choose per test or per interview yourself, and what the default is.
What do reviewers see?
The signal is in how the candidate used the AI, not just the final code. Look for the full conversation, a replay of the session, and whether you can watch live.
Platform by platform
| Platform | Kind of AI | Who turns it on | What reviewers see |
|---|---|---|---|
| CoderPad | AI Assist: chat with frontier models (Claude, GPT, Gemini), plus Edit Mode that writes and changes code and Plan Mode for planning in interviews | Optional, opt-in per company | The AI conversation in the pad's playback |
| HackerRank | AI assistant in the IDE: chat, agent mode and inline completions (some features in a paid AI add-on) | On by default in its interview workflows; turning it off goes through HackerRank support | AI conversations in the interview report; interviewers can watch them live |
| CodeSignal | Cosmo, an assistant built into the IDE | Enabled for the account, then on or off per assessment | The full transcript and a session replay in the report |
| Codility | Cody, a guardrailed assistant that helps clarify the task but won't write a full solution; a separate AI Copilot in VS Code for interviews | On or off per test or interview; Cody isn't available on custom content | Every prompt and response, read-only |
| Kendor | Claude Code and Codex, running in the terminal of a full sandbox with the real project, database and dev server | Off by default; on per screen or live interview, and flippable during a live session | The agent transcript turn by turn, a terminal replay, the code diff, and the live terminal while the attempt runs |
CoderPad
CoderPad started with a ChatGPT window in each pad and has grown it into AI Assist. In 2026 it added an Edit Mode that lets the AI write and change code, and a Plan Mode for working through an approach before coding. The conversation is saved in the interview playback. It suits live pair-programming interviews where you want to see how someone prompts and checks AI output.
HackerRank
HackerRank's IDE has an AI assistant with chat, an agent mode and inline completions. Its help centre says the assistant is enabled by default in its interview workflows, including code repository rounds, and that you contact support to turn it off. Interviewers see the AI conversations live and in the report. Some AI features need its AI add-on.
CodeSignal
CodeSignal's AI-assisted assessments and interviews are built around Cosmo, an assistant inside the IDE. An admin enables it for the account, then it can be turned on or off per assessment. Reviewers get the full transcript of the candidate's AI conversation and a session replay.
Codility
Codility's Cody is guardrailed on purpose: it helps candidates understand the task and iterate, but won't produce a full solution, and it doesn't affect scoring. Test creators switch it on per test, and interviewers per interview. Every prompt and response is logged. Codility also offers an AI Copilot in VS Code for interviews, with completions and chat.
Kendor
Kendor gives candidates the same agents many engineers use at work: Claude Code and Codex, run from the terminal against the real project. That's a full development environment in the browser, with the company's own repository, databases and a running app. The agents can run on your organization's Anthropic or OpenAI account, with token budgets and keys that never enter the sandbox, or on the candidate's own account. Reviewers read the agent conversation turn by turn next to the code, replay the terminal, and can watch live.
Which fits which interview?
- You want to see AI fluency in a short live interview: a chat assistant in a shared pad works, as with CoderPad AI Assist or HackerRank's IDE assistant.
- You want AI help without handing over the answer: a guardrailed assistant like Codility's Cody.
- Your engineers work with coding agents on a real codebase: test that directly. Give candidates a real project and a real agent, and review how they directed it. This is what Kendor is built for.
- You want no AI at all: every platform above lets you turn it off. Check how: per test, per interview, or through support.
Doing this in Kendor
Turn on AI assistants in the terminal in a screen's settings, choose whose account the agents run on, and set a token budget. Candidates are told before they start what's allowed and what's recorded. Our guide Should you allow Claude Code in technical interviews? covers how to design the task, set the budget and review the transcript, and Turn on Claude Code and Codex has the steps.
Sources
- CoderPad: AI policy, 2026 product recap, ChatGPT integration launch
- HackerRank: Interview features, AI features
- CodeSignal: AI-assisted coding assessments and interviews
- Codility: Meet Cody, AI Copilot in interviews, AI on or off