Self-hosted AI assistant for developers
Run your own AI agents on your laptop or server with one Go binary. Any model, local or hosted, reachable from the terminal, the web or your chat apps.
One command to install
A single binary in ~/.local/bin, a setup wizard on first run, SQLite under ~/.fastclaw. No Node, Python or Docker required.
Your model, your keys
OpenAI, Anthropic, DeepSeek, OpenRouter or a local Ollama model. Switch per agent, or per conversation with /model in fastclaw chat.
Agents with real memory
Each agent has a SOUL.md personality, IDENTITY.md and a long-term MEMORY.md it keeps up to date between conversations.
Reminders and recurring jobs
Ask "remind me at 9 every morning" and the agent creates a cron job. Results arrive in the web chat or the IM channel you asked from.
Scriptable from the CLI
fastclaw agents init, config, files and rm drive the same store as the dashboard. A running gateway hot-reloads on change.
A personal agent you actually own
FastClaw installs like a CLI tool and runs like a service. Run fastclaw, pick a provider in the setup wizard and you have a working agent with a web dashboard on localhost:18953. Everything — agents, sessions, settings — lives in one SQLite file you can back up or move.
Use a hosted model like OpenAI or keep it local with Ollama. Each agent can use a different model, so a cheap fast model can triage while a stronger one handles real work.
Talk to it from anywhere
The same agent answers in the terminal (fastclaw chat), the browser, and your phone. Bind it to Telegram, or on a Mac let it answer iMessage direct messages. Pair the channel with /pair <code> so only you get owner access.
Let it work while you don't
Agents create their own scheduled jobs — daily summaries, reminders, recurring checks — and push the results back into the chat you started from. Install skills from ClawHub or GitHub to give them new abilities, or write your own SKILL.md.
Works with
All Integrations
Ollama
Point FastClaw at Ollama to run AI agents on local, open-weight models — no API key, no data leaving your machine.
Telegram
Turn any FastClaw agent into a Telegram bot in minutes: paste a BotFather token, pair your account, and chat in DMs or groups.
iMessage
Run FastClaw on a Mac and let an agent answer iMessage direct messages sent to that Mac's Apple ID.
Frequently asked questions
More Use cases
All Use casesSaaS builders
Give every customer their own AI agent without building a runtime. FastClaw handles agents, end-user isolation, sandboxed tools and usage metering over an API.
Teams
Deploy shared AI agents inside the chat tools your team already uses. Self-hosted, with per-person sessions, shared skills and scheduled reports.
Communities
Run an AI agent for your community on Telegram, Discord or LINE. Every member gets their own session, and untrusted chatters stay inside a sandbox.
Try it on your own agents
Start free in FastClaw Cloud, or self-host the single binary in a minute.