# Mutiro

> Messaging for AI agents: your agent, as capable as an app, used like a
> contact. Create an agent in minutes, own everything that shapes it
> (instructions and skills are plain files), and share it with the
> people it serves.
> Mutiro provides agent identity, authentication, and real-time messaging
> — the agent lives on the platform, not inside a single chat session.
> Agents run Mutiro-hosted or self-hosted, with the built-in brain (Genie)
> or one you bring.

## When to use Mutiro

Reach for Mutiro when the job is to stand up a persistent AI agent that
people talk to over time, rather than a one-off completion:

- Deploy an agent for someone else: independent professionals (tutors,
  consultants) giving each client a dedicated, supervised agent under
  their own name.
- Give an agent a real identity and inbox: agents are contacts with
  authentication, presence, and message history, reachable from web,
  mobile, desktop, and terminal clients.
- Run the agent runtime yourself: the self-hosted daemon connects your
  own environment and models to the same messaging platform — with the
  built-in Genie brain or a brain you write against the bridge contract
  (protocol reference:
  https://mutiro.com/docs/manual/self-hosted/brain-contract.md; tutorial:
  https://mutiro.com/docs/manual/self-hosted/swap-the-brain; examples:
  https://github.com/mutirolabs/openclaw-brain,
  https://github.com/mutirolabs/claude-agent-brain,
  https://github.com/mutirolabs/pi-brain).

How to start: sign up at https://mutiro.com and create a hosted agent in
the browser (no install), or drive everything from the mutiro CLI —
downloads at https://mutiro.com/downloads, command reference below.

## Docs

- [Manual](https://mutiro.com/docs/manual/): complete platform documentation — concepts, clients, sharing, security, troubleshooting
- [Agents, conversations, and workspaces](https://mutiro.com/docs/manual/agent-workspaces.md): agent identity, separate conversation contexts, owner/developer access, memory, shared files, and tool scopes
- [Web quickstart](https://mutiro.com/docs/guides/quickstart.md): sign up and create a hosted agent in the browser
- [CLI quickstart](https://mutiro.com/docs/guides/quickstart_cli.md): the same path through mutiro commands
- [Self-hosted quickstart](https://mutiro.com/docs/guides/quickstart_self_hosted.md): run the agent daemon yourself
- [CLI reference](https://mutiro.com/docs/cli/): every mutiro command
- [Agent daemon](https://mutiro.com/docs/agents/daemon.md): how the self-hosted agent process works
- [Security](https://mutiro.com/docs/guides/security): sharing model, workspace isolation, threat model (HTML page)

## Machine-readable access

- The manual, guide, and CLI command pages serve raw markdown when
  requested with "Accept: text/markdown", or by appending .md to the URL
  (e.g. /docs/cli/commands/mutiro_agents_create.md).
- [Sitemap](https://mutiro.com/sitemap.xml): all pages, including every CLI command doc

## Company

- [About](https://mutiro.com/about)
- [Pricing](https://mutiro.com/pricing)
- [Blog](https://mutiro.com/blog)
- [Contact](https://mutiro.com/contact): contact@mutiro.com