Skip to main content
Peeps is autonomous verification for AI-native software teams: AI agents that write, run, and self-heal your end-to-end regression suites, with AI failure analysis on every PR and release. Point Peeps at your app. It maps the product, writes test cases in plain English, generates the E2E code, then runs and self-heals the suite on every change. QA becomes supervision — engineers review verdicts instead of clicking through flows.

​
Start here

Keep tests in your repository

Run your own Playwright suite in your CI, with Peeps analyzing, fixing and adding tests.

Connect GitHub

Trigger runs from CI and get failure analysis and healing on every change.

Use your coding agent

Troubleshoot, run, and fix tests from Claude Code, Cursor, or Codex.

Set up SSO

Let your people sign in with your company login through Okta.

Troubleshoot and fix a failure

Find out why a test fails and publish a verified fix.

Run tests

Approve a plan, then follow the run to its result.

Create and import tests

Cover real gaps, or bring in your Playwright tests.

Command reference

Every MCP command, grouped by the job you want done.

​
Where Peeps reaches you

Peeps runs on its own. The suite executes on every change, a failure is analyzed and given a cause and a confidence level, a script issue is healed and re-verified, and you hear about it in Slack or email. Peeps has five surfaces: Slack, email, Peeps AI chat, the Peeps web app, and the MCP server. These docs cover tests that live in your own repository, the MCP server, the GitHub and GitLab integrations, and single sign-on.

​
Scope today

Peeps verifies pre-production — end-to-end, UI, and workflow tests, run as Web (Playwright). Everything in these docs is scoped to that.