AI agents in the terminal
The AI agents engineers actually use, in the sandbox. On your keys, or on theirs.
If the job involves working with an agent, screen for it. Candidates run Claude Code, Codex, Antigravity, GitHub Copilot, Cline, OpenCode or Cursor in their sandbox terminal. Bill it to your organization's Anthropic or OpenAI key with a token budget, or let candidates sign in to their own account so model usage costs you nothing.
01
Their own agent, on their own account.
Engineers don't all use the same agent. Set a screen to the candidate's own account and they sign in to their Claude, ChatGPT, Google, GitHub Copilot, Cline or Cursor account in the sandbox and work with the tool they use every day. Your organization needs no API key and sets no budget, and model usage costs you nothing.
- Claude Code, Codex, Antigravity, GitHub Copilot, Cline, OpenCode and Cursor, in one terminal
- OpenCode's free models need no account, so candidates without a paid plan aren't left out
- Sign-ins are deleted when the candidate submits, and your organization never sees them
02
Your key never enters the sandbox.
Claude Code and Codex in the sandbox talk to Kendor's gateway with a session token that opens nothing on its own. The gateway checks the session and its budget, adds your organization's key and forwards the call to Anthropic or OpenAI. There's nothing in the environment for a candidate to print or copy.
- Keys are encrypted at rest, and only organization owners can change them
- Anthropic can connect through identity federation, with no long-lived key at all
- The gateway only reaches Anthropic and OpenAI. It isn't a way out to the internet
$ env | grep -E 'ANTHROPIC|OPENAI'ANTHROPIC_BASE_URL=http://kendor-gateway:4102/anthropicANTHROPIC_AUTH_TOKEN=eyJhbGciOiJIUzI1NiIs…OPENAI_BASE_URL=http://kendor-gateway:4102/openai/v1OPENAI_API_KEY=eyJhbGciOiJIUzI1NiIs…a session token, not a vendor key03
A budget per candidate, a cap per month.
Give each screen a token budget per candidate, counted across all its challenges, and set a monthly cap for the whole organization. Candidates see a ring in their header fill toward the budget, and the gateway stops the agent as soon as either limit is reached.
- Counts what the agent reads and writes, not cached context
- Usage is billed to your own Anthropic or OpenAI account
04
Off until you turn it on.
AI assistants stay off for every screen until you switch them on in the builder. Candidates get the agents your organization has keys for, or all seven on their own account, and their start page tells them before they begin where their work goes and that reviewers can see their terminal.
05
Watch the terminal live, read-only.
Open the Terminal tab on a candidate's review page while they work and you see their shell as it happens: what they ask the agent, what it runs and what they keep. Reviewers can watch, but they can't type into it.
06
Read every prompt afterwards.
The AI agents tab keeps the conversation in order: each prompt, the tools the agent called and what it answered, with the time and model on every turn. You judge how someone works with an agent, not only the code it left behind.
- Candidates are told up front that the conversation is saved for reviewers
- PromptInvoices with a zero total are still being emailed. Find out why and fix it with a test.Tool · Read
{"file_path":"/workspace/internal/invoice/send.go"}▸ Tool output - AgentThe guard checks inv.Total < 0 instead of inv.Total <= 0, so zero-total invoices get through.Tool · Edit
{"file_path":"/workspace/internal/invoice/send.go","old_string":"if inv.Total < 0 {","new_string":"if inv.Total <= 0 {"}▸ Tool output
07
Switch agents on mid-interview.
In a live interview or live coding session, the host turns AI agents on or off for everyone from the room's top bar. The running sandbox picks up the change straight away, without a restart.
- Everyone in the room sees whether AI is on
- Same organization keys and gateway as your screens
How it works
Two settings, then it's in their terminal.
Choose who pays
Connect an Anthropic or OpenAI key under Integrations once for the organization, with a monthly cap if you want one. Or skip it and let candidates use their own account.
Turn it on for a screen
Switch on AI assistants in the screen builder and give it a token budget per candidate.
Candidates run it
They find claude, codex or the other agents ready in their sandbox terminal, and you can watch from the review page.
Questions
Who pays for model usage?
You choose per screen. On your organization's keys, you pay Anthropic or OpenAI directly, and Kendor doesn't resell tokens or add a margin. On the candidate's own account, candidates use their own plan and model usage costs you nothing.
What does connecting without a key mean?
For Anthropic, your Claude Console organization can trust Kendor through workload identity federation, so no long-lived API key is created or shared. OpenAI uses an API key.
Doesn't allowing AI hide what the candidate can do?
No. Reviewers read the agent transcript beside the code: what the candidate asked, what they accepted, what they rewrote. That is the part your reviewers judge.
Which agents can candidates use?
With your organization's keys: Claude Code and Codex. With their own accounts: Claude Code, Codex, Cursor, GitHub Copilot, Cline, OpenCode and Antigravity. OpenCode's free models need no account.
More of Kendor
All features- Coding screensAsync take-homes in a real browser IDE, reviewed by your team.Read more
- Live codingOne link, one shared sandbox, a verdict from every interviewer.Read more
- InterviewsSelf-booking, invites that update in place, live review.Read more
- AI review summaryA first read of each attempt, with evidence. Never a score.Read more
- Real environmentsSeven pinned languages, four databases, one kendor.yaml.Read more
- Repo importCloud or self-hosted Git, private repos too, one fixed snapshot.Read more
- Ashby integrationAshby stages send screens or book live coding; results come back as fields.Read more
- Greenhouse integrationGreenhouse stages send screens and live coding; notes come back.Read more
- Lever integrationLever stages send screens and live coding; each step comes back as a note.Read more
- EnterpriseHackathonsEvent page, applications and a build day in real sandboxes.Read more
Run your next screen on Kendor. Read what the code says.
Create your workspace, import a repository or start from a template, and send your first screen today.
Hiring at scale? Talk to us