# TaloTrace Documentation

## Instructions for AI Agents

- For clean Markdown of any page, append `.md` to the page URL
- For section-specific indexes, append `/llms.txt` to any section URL
- For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.talotrace.ai/_mcp/server

## Docs

- [Welcome to TaloTrace](https://docs.talotrace.ai/welcome.md): Test your web or Android app with an AI agent, from your coding agent, the web app or CI.
- [Test localhost with Claude Code](https://docs.talotrace.ai/quickstart/web-local-claude-code.md): From a web app on your laptop to a first verdict in about ten minutes.
- [Test a hosted URL](https://docs.talotrace.ai/quickstart/web-hosted.md): Test a deployed web app from the TaloTrace web app, with no install.
- [One-command setup](https://docs.talotrace.ai/install/setup-command.md): talotrace setup signs you in and prepares your coding agent in one step.
- [Manual install for Claude Code](https://docs.talotrace.ai/install/claude-code.md): Install the TaloTrace plugin from its marketplace when you do not use setup.
- [Manual install for Cursor, Codex and others](https://docs.talotrace.ai/install/cursor-codex-other.md): Add the TaloTrace MCP server by hand to any MCP client.
- [Sign in](https://docs.talotrace.ai/sign-in.md): One browser sign-in, shared by the CLI and every agent on this computer.
- [API keys](https://docs.talotrace.ai/api-keys.md): Let a cloud agent or CI act in one project with no browser.
- [Check your install](https://docs.talotrace.ai/check-your-install.md): talotrace doctor and the diagnose tool say what is missing.
- [How TaloTrace works](https://docs.talotrace.ai/concepts/how-it-works.md): From a project to a verdict and its evidence.
- [Glossary](https://docs.talotrace.ai/concepts/glossary.md): The words TaloTrace uses, in one line each.
- [Projects and apps](https://docs.talotrace.ai/concepts/projects-and-apps.md): The project is the unit of testing. An app is the product under test.
- [Scenarios and test plans](https://docs.talotrace.ai/concepts/scenarios-and-test-plans.md): A scenario is one test. A test plan runs several together.
- [Runs and verdicts](https://docs.talotrace.ai/concepts/runs-and-verdicts.md): A run passed only when it ended, executed to the end and its analysis found no failure.
- [Cards](https://docs.talotrace.ai/concepts/cards.md): The questions TaloTrace needs you to answer.
- [Evidence](https://docs.talotrace.ai/concepts/evidence.md): What a run recorded, kept as one bundle per attempt.
- [Credits](https://docs.talotrace.ai/concepts/credits.md): Runs spend credits. You can see the estimate first and cap it.
- [Test an app on localhost](https://docs.talotrace.ai/web/test-localhost.md): Run TaloTrace against a web app on your laptop, with no change to the app.
- [Test a deployed URL](https://docs.talotrace.ai/web/test-deployed-url.md): Point TaloTrace at a public HTTPS page.
- [Answer cards](https://docs.talotrace.ai/guides/answer-cards.md): Read the questions TaloTrace asked and answer the ones that are yours.
- [Read a run](https://docs.talotrace.ai/guides/read-a-run.md): From the verdict down to the step that failed.
- [Prove a fix](https://docs.talotrace.ai/guides/prove-a-fix.md): Rerun once and compare.
- [Project health](https://docs.talotrace.ai/guides/project-health.md): The Health blocks on Home show what needs you and what is failing.
- [Live phone](https://docs.talotrace.ai/mobile/live-phone.md): Try an Android change on a cloud phone that runs your working tree.
- [MCP tools](https://docs.talotrace.ai/reference/mcp.md): Every tool the TaloTrace MCP server offers your coding agent.
- [Setup and diagnose](https://docs.talotrace.ai/reference/mcp/setup-diagnose.md): Arguments of each tool, as the MCP server declares them.
- [Scenarios, knowledge and test data](https://docs.talotrace.ai/reference/mcp/scenarios-knowledge-test-data.md): Arguments of each tool, as the MCP server declares them.
- [Runs and evidence](https://docs.talotrace.ai/reference/mcp/runs.md): Arguments of each tool, as the MCP server declares them.
- [Requests and the live phone](https://docs.talotrace.ai/reference/mcp/requests-dev.md): Arguments of each tool, as the MCP server declares them.
- [CLI commands](https://docs.talotrace.ai/reference/cli.md): Every talotrace command and flag.
- [Exit codes](https://docs.talotrace.ai/reference/cli/exit-codes.md): What the process exit code says.
- [Problems](https://docs.talotrace.ai/reference/problems.md): Every problem code, what it means and what to do.
- [Configuration](https://docs.talotrace.ai/reference/configuration.md): Packages, requirements and environment variables.
- [Limits](https://docs.talotrace.ai/reference/limits.md): What TaloTrace caps, and where.
- [Troubleshooting](https://docs.talotrace.ai/help/troubleshooting.md): Find your symptom, then the code and the fix.
- [Changelog](https://docs.talotrace.ai/changelog.md): What changed in the CLI, the plugin and the API, by date.