---
title: "What Bugbearer does"
description: "AI agents test your web app in a real browser, find bugs, and record each run as a test case that replays without a model."
---

> Documentation Index
> Fetch the complete documentation index at: https://docs.bugbearer.com/llms.txt
> Use this file to discover all available pages before exploring further.

# What Bugbearer does

Bugbearer tests a web app with AI agents. An agent opens your app in a real browser, uses it like a person, and reports the bugs it finds. Each agent run is saved as a test case. Later runs replay that test case without a model.

## How it works

1. **Connect an app**

   Give Bugbearer a GitHub repo or a staging URL. See [Connect an app](https://docs.bugbearer.com/connect-an-app).
2. **Let setup finish**

   For a repo, a setup agent works out how to start the app in Docker and how a test user signs in. You review the plan and add any test keys. See [Secrets and test keys](https://docs.bugbearer.com/secrets-and-test-keys).
3. **Run tests**

   Agents run test cases in a browser on a [runner](https://docs.bugbearer.com/runners). Each run records video, screenshots, console output and network events.
4. **Read the bugs**

   Findings that hold up appear on the Bugs board. Each bug has steps, evidence and a downloadable bundle. See [Bugs and bug bundles](https://docs.bugbearer.com/bugs-and-bug-bundles).

## What you need

- A Bugbearer account. You sign in with GitHub or Google.
- An app to test: a GitHub repo, or a staging or preview URL that is not production.

## Use it from a coding agent

A coding agent can connect an app and start setup for you through the Bugbearer MCP server. See [Connect the MCP server](https://docs.bugbearer.com/onboarding-mcp) and [Agent-driven onboarding](https://docs.bugbearer.com/agent-onboarding).

> **Caution**
>
> Agents sign up, change settings and submit forms in the app they test. Point Bugbearer at a staging or preview app, never at production.

Source: https://docs.bugbearer.com/index.mdx