Skip to content

OtterBot Docs

Everything you need to run, configure, and extend OtterBot — the free, open-source, local-first multi-agent AI workspace.

Quick Start API Reference GitHub

What Is OtterBot?

OtterBot is built around agent profiles. Instead of one assistant with a handful of modes, you run a roster of independent agents — each with its own personality, model, memory, skills, credentials, and scheduled tasks. A COO agent coordinates the team, agents talk to each other over a message bus you can watch live, and trusted agents can spawn subagents for focused work.

Everything runs on your own machine, and every database is encrypted at rest.

Explore the Docs

  • Getting Started — Install the CLI, run the daemon, and configure your first agent.
  • Architecture — Agent profiles, the orchestrator, runtimes, the message bus, and storage.
  • Agent System — Agent roles, the COO, subagents, and Agent Studio.
  • Features — Roster, chat, memory, skills, the scheduler, integrations, and the 3D scene.
  • API Reference — REST endpoints and Socket.IO events.

At a Glance

Aspect Detail
Local-first Runs on your own machine — no cloud required
Multi-agent A roster of independent agent profiles
Encrypted Every SQLite database is encrypted at rest
Model providers Anthropic, OpenAI, LM Studio, Ollama
Per-agent Personality, model, memory, skills, schedule, credentials
Observable A live message bus and a 3D agent scene

Quick Start

Install the CLI from npm, start the daemon, and open the web UI:

npm install -g otterbot
otterbot start
otterbot web

The daemon serves the web UI on http://localhost:3001. On first run, a setup wizard configures your first agent. See the Getting Started guide for CLI commands, running from source, and configuration.