AI onboarding
Bird is built to be driven by agents as much as by humans, and setting that up is one paste. This page covers the three ways in: a setup prompt that installs the bird CLI and Bird's skills, the MCP server for clients that call tools instead of running a shell, and the machine-readable docs corpus for agents that prefer reading. The CLI is the default because it is the leanest agent surface: JSON output, semantic exit codes, far fewer tokens per operation than tool calls, and a self-describing command tree. Set it up once and your agent can send, inspect, and debug on Bird without you copy-pasting anything.
1. Set up your agent (CLI + skills)
Paste one prompt into your coding agent and it does the rest: it installs the bird CLI, signs you in with bird auth login, installs Bird's skills, and proves the setup works by sending a test email to your own inbox. On the home page (and the Products and Email pages), the Claude Code / Codex / Cursor buttons copy the prompt tailored to each client; set up your coding agent walks through the same flow. The skills teach the agent Bird's workflows (send and confirm delivery, find a verified sender, manage webhook endpoints) so it reaches a correct command on the first try.
2. Connect the MCP server (when it fits)
If your client calls tools instead of running a shell (Claude Desktop, ChatGPT developer mode, mobile), point it at Bird's hosted MCP server instead. It is a URL and a browser sign-in, nothing to install:
Codebeispiel
https://mcp.bird.comThe first time a client connects, your browser opens a Bird consent screen: you sign in, pick the workspace and the permissions to grant, and approve. No API key to create, no secret in your client config. The endpoint, the full tool surface, the delegated OAuth model, and the local-stdio option (bird mcp) are in the MCP server guide.
3. Feed your agent the docs corpus
Everything on this docs site is also published in machine-readable form, so agents that prefer reading over tool calling get the same content you're looking at now:
- /llms.txt: the index, a flat, link-rich summary of every page, a starting point the agent can follow links from.
- /llms-full.txt: the full corpus, every doc page concatenated as Markdown, for agents that want the whole thing in context.
Both are regenerated with every deploy of the site, so they match the published docs. Point your agent at the index URL, paste the full corpus into a long-context session, or use the AI tools in the sidebar of any docs page: Copy as Markdown copies that page's source, and Open llms-full.txt opens the corpus. The dashboard's onboarding page mirrors the same Copy-for-AI shortcuts.
The formats, sizing, and retrieval tips are covered in the llms.txt guide.
4. Try it
Once your agent is set up, give it a real task end to end:
Find my email with whoami, then send me a test email from onboarding@messagebird.dev and tell me when it's delivered.
The agent reads your address with whoami, sends from Bird's shared onboarding domain (which delivers to verified members of your workspace, so your own inbox works with no domain to verify), and polls the status until it lands, so the proof shows up in your inbox. That's the same happy path as the send-your-first-email quickstart, driven by the agent. For a deterministic check with no real inbox, send to delivered@messagebird.dev instead; see the testing sandbox.
Next steps
- Self-serve signup: for an agent that needs its own Bird account rather than yours. Create the account, organization, and workspace from the terminal, no browser.
- CLI for agents: the bird CLI's agent contract. JSON output, semantic exit codes, OAuth login.
- Agent skills: the bird-ai marketplace plugin and what each client gets.
- MCP server: the hosted endpoint, the full tool surface, the delegated OAuth model, and the local-stdio option.
- llms.txt & Markdown docs: how the machine-readable corpus is built and how to retrieve from it.
- Send your first email: the human-driven version of the happy path above.