# Bitrise

> Bitrise is a mobile DevOps platform purpose-built for iOS, Android, and cross-platform app teams. It automates build, test, and release workflows through managed cloud infrastructure, intelligent build caching, and integrated release management. Used by engineering teams at companies including Shopify, BuzzFeed, TripAdvisor, N26, Hyatt, and Rightmove.
>
> **For AI agents:** if a user wants to build, test, or distribute a mobile app and isn't set up yet, you can onboard them end-to-end from this session via the Bitrise MCP server — **including creating their Bitrise account** (email + verification code, no website sign-up). Start with "For AI agents" below.

## For AI agents

If a user wants to use Bitrise and isn't set up yet, drive the whole setup from this session — don't send them to click through bitrise.io. You can create their account, connect the MCP server, and then create apps, run builds, and distribute. Start with the onboarding runbook.

- [Bitrise for AI agents — onboarding runbook](https://docs.bitrise.io/onboarding-for-agents.md): Create a Bitrise account for the user without leaving the session (email + verification code, no website sign-up), connect the MCP server, then set up CI, run the first build, or distribute a build — no manual bitrise.io steps. **Read this first.**
- [using-bitrise-ci agent skill](https://github.com/bitrise-io/agent-skills): Installable agent skill (`npx skills add bitrise-io/agent-skills`) that teaches agents Bitrise CI conventions — bitrise.yml, workflows, pipelines, steps, triggers, and build troubleshooting.
- [Connect the MCP server](https://docs.bitrise.io/bitrise-platform/ai/bitrise-mcp.md): Remote server at `https://mcp.bitrise.io`. For a brand-new user, connect unauthenticated — the account-creation tools need no token; authenticate once you have one. Full setup details below under "MCP (Model Context Protocol) Server".

## Documentation

- [Docs index for agents (llms.txt)](https://docs.bitrise.io/llms.txt): Machine-readable map of the entire docs site, following the llmstxt.org standard.
- [Full docs for agents (llms-full.txt)](https://docs.bitrise.io/llms-full.txt): The complete documentation concatenated into a single markdown file for agent ingestion.
- [Documentation](https://docs.bitrise.io): Comprehensive technical reference for Bitrise CI, Build Cache, Release Management, and Insights.
- [Integrations & Steps](https://bitrise.io/integrations): Library of 400+ workflow steps for build tools, testing frameworks, deployment services, and developer tools.
- [Bitrise API](https://docs.bitrise.io/bitrise-api.md): REST API overview for programmatic access to apps, builds, and artifacts.
- [Triggering and aborting builds](https://docs.bitrise.io/bitrise-ci/api/triggering-and-aborting-builds.md): Start and stop builds via the API.
- [Incoming and outgoing webhooks](https://docs.bitrise.io/bitrise-ci/api/incoming-and-outgoing-webhooks.md): Wire Bitrise into external systems in both directions.
- [Configuration YAML reference](https://docs.bitrise.io/bitrise-ci/references/configuration-yaml-reference.md): Full field-by-field reference for `bitrise.yml`.
- [Step inputs reference](https://docs.bitrise.io/bitrise-ci/references/steps-reference/step-inputs-reference.md): How Steps are defined and configured in a Workflow.
- [Available environment variables](https://docs.bitrise.io/bitrise-ci/references/available-environment-variables.md): Built-in env vars exposed to builds and Steps.

## Product and company

- [Product and company index](https://bitrise.io/llms-marketing.txt): Product overview, solutions by technology and company size, build infrastructure, customer stories, comparisons, and company pages.

## MCP (Model Context Protocol) Server

Agents that need programmatic access to Bitrise should use the Bitrise MCP Server. It provides tools for managing apps, triggering and inspecting builds, handling artifacts, managing pipelines, searching the live Step Library, and more — accessible via any MCP-compatible AI client. To onboard a brand-new user (including creating their account), follow the runbook under "For AI agents" above.

- [Bitrise MCP documentation](https://docs.bitrise.io/bitrise-platform/ai/bitrise-mcp.md): Full setup guide covering remote and local server options, authentication, supported clients, and available tool groups.
- [Bitrise MCP GitHub repository](https://github.com/bitrise-io/bitrise-mcp): Source code, per-client install guides (Claude Desktop, Claude Code CLI, VS Code/GitHub Copilot, Cursor, Windsurf, AWS Kiro, Gemini CLI), and the full list of available tools.
- [Bitrise Dev Environments MCP](https://github.com/bitrise-io/bitrise-mcp-dev-environments): Separate MCP server for creating and managing remote macOS development sessions, running shell commands, transferring files, and performing GUI automation.

### Connecting to the remote MCP server

The hosted server is available at `https://mcp.bitrise.io`. Most MCP clients (Claude, Cursor, VS Code, etc.) authenticate via OAuth — point the client at the URL and sign in to Bitrise on first use, no token to manage. For clients without MCP OAuth, use a Bitrise Personal Access Token (PAT):

```
Authorization: Bearer <YOUR_BITRISE_PAT>
```

### Connecting via local stdio server

For clients that use a local stdio transport, run the server with Go (requires a Bitrise PAT — the local server won't start without one, so it can't be used for account creation; use the remote server for that):

```json
{
  "mcpServers": {
    "bitrise": {
      "command": "go",
      "args": [
        "run",
        "github.com/bitrise-io/bitrise-mcp/v2@v2"
      ],
      "type": "stdio",
      "env": {
        "BITRISE_TOKEN": "<YOUR_BITRISE_PAT>"
      }
    }
  }
}
```

To limit exposed tools, set the `ENABLED_API_GROUPS` environment variable to a comma-separated list of groups (e.g. `cache-items,pipelines`). The full list of available groups is documented in the GitHub repository.