Run your own AI agents.One binary. Any LLM.
FastClaw is an agent factory: create, manage and run AI agents — each with its own personality, memory, skills and tools — and put them on the web, your API, or nine chat apps.
$ curl -fsSL https://raw.githubusercontent.com/fastclaw-ai/fastclaw/main/install.sh | bashmacOS · Linux · Windows. Drops one binary into ~/.local/bin.
Telegram
Discord
Slack
Feishu
WeCom
WeChat
WhatsApp
LINE
iMessageSee it in action.One console for every agent.
Chat with your agents, shape their personality, connect channels and schedule work — all from the FastClaw console.
Agents answer with rich markdown and code — here a Code Reviewer catches a SQL injection and a race condition before merge.
Everything an agent needs,built into the runtime.
FastClaw handles model calls, tool execution, sandboxing, sessions and channels — so you can spend your time on what your agents do.
Agents
Create and run many AI agents from one FastClaw install — each with its own SOUL.md personality, memory, skills, model and sharing settings.
Models
Run FastClaw agents on OpenAI, Anthropic, DeepSeek, Ollama, OpenRouter, Groq, Mistral or any OpenAI-compatible API, with per-agent overrides and retries.
Channels
Bind any FastClaw agent to Telegram, Discord, Slack, Feishu, WeCom, WeChat, WhatsApp, LINE or iMessage — with per-chat session isolation and owner pairing.
Tools & Sandbox
Built-in exec, file and web tools, run in an E2B cloud sandbox or a Docker container, with automatic skill and workspace hydration and file sync.
Skills & MCP
Install skills from ClawHub, skills.sh or GitHub, connect MCP servers, add JSON-RPC plugins and let agents read Notion, GitHub and Linear.
Runtime API
Drive FastClaw agents from any app: OpenAI-compatible chat completions, agent management, per-end-user isolation, usage reporting and ACP.
Scheduler
FastClaw agents schedule their own jobs — one-off reminders, intervals or cron expressions — and deliver results to the chat that asked.
One directory.Everything your agents need.
No config file to mount. The database is the source of truth; skills live on disk next to it.
- ~/.fastclaw/Your FastClaw home
- ├── fastclaw.dbUsers, agents, sessions, keys, configs
- ├── skills/Shared skills — bundled + installed
- ├── agents/One folder per agent
- │ └── <agentId>/agent/skills/Agent-private skills
- └── users/Per-account data
- └── <uid>/skills/Personal skills
Env-only bootstrap
Port, bind, storage and sandbox come from FASTCLAW_* env vars. Everything else is edited live in the console.
SQLite to Postgres
Start on one SQLite file. Point FASTCLAW_STORAGE_DSN at Postgres when you go multi-pod.
Laptop to Kubernetes
The same binary runs in the foreground, as a launchd / systemd daemon, in Docker or on k8s.
Built for people whoship agents to real users.
From a personal assistant on your Mac to the agent backend of a multi-tenant SaaS.
SaaS 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.
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.
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.
Agencies
Build an agent once, fork it for each client, and deliver it on WeChat, WhatsApp or the web — all from one FastClaw deployment with per-client quotas.
Running in under a minute
Install the binary, run the setup wizard, and open the dashboard. No config files — bootstrap is env-only.
- 01
Install
One command drops the fastclaw binary into ~/.local/bin and adds it to your PATH.
$ curl -fsSL https://raw.githubusercontent.com/fastclaw-ai/fastclaw/main/install.sh | bash - 02
First run
Launch fastclaw to open the setup wizard — configure an LLM provider and it creates your first agent.
$ fastclaw - 03
Open the dashboard
Manage agents, skills, models, channels, and API keys at http://localhost:18953.
$ open http://localhost:18953
Bring any model.Reach every channel.
OpenAI, Anthropic, DeepSeek, Ollama, OpenRouter and any OpenAI-compatible API — with E2B or Docker sandboxes, Postgres and MCP.
OpenAI
Run FastClaw agents on OpenAI models — set a deployment-wide default, override per agent, and expose your agents through an OpenAI-compatible API.
Anthropic
Run FastClaw agents on Claude through Anthropic's native Messages API — with prompt caching, extended thinking and image input preserved.
DeepSeek
Run FastClaw agents on DeepSeek — a built-in preset, thinking-mode support and per-agent model overrides for low-cost, capable agents.
Ollama
Point FastClaw at Ollama to run AI agents on local, open-weight models — no API key, no data leaving your machine.
OpenRouter
Connect FastClaw to OpenRouter and give every agent access to models from many vendors with a single API key.
E2B
Run every FastClaw agent command in an isolated E2B cloud sandbox — skills and workspace hydrated in, files synced back after each call.
Docker
Use Docker as FastClaw's sandbox backend to isolate agent commands on your own hardware — or deploy FastClaw itself with Docker Compose.
PostgreSQL
Swap FastClaw's default SQLite file for PostgreSQL to run multiple pods behind a load balancer, with automatic schema migrations.
MCP
Give FastClaw agents external tools through the Model Context Protocol — add http or stdio MCP servers per agent from the console.
Weighing your options?
Honest, side-by-side comparisons — including when another tool is the better fit.
FastClaw vs OpenClaw
OpenClaw is a personal AI assistant you run yourself. FastClaw is a Go agent runtime for many users and apps. Compare install, multi-tenancy, API, license.
OpenClaw alternative
Looking for an OpenClaw alternative? FastClaw is a single-binary Go runtime for serving agents to many users and apps — and it reads OpenClaw skills.
FastClaw vs Dify
Dify is a visual platform for LLM workflows and RAG apps. FastClaw is a lightweight runtime for autonomous agents in chat apps and products. Compare the two.
From the blog
All postsDeploy your own Telegram AI bot in 15 minutes
Step-by-step: install FastClaw, connect a model provider, give your agent a personality, bind it to a Telegram bot and pair it with your account — no public URL or webhook required.
How to add AI agents to your SaaS without building an agent platform
A practical guide to giving every customer their own AI agent: user-scope API keys, per-user agents, end-user isolation, streaming, files and usage-based billing with the FastClaw /v1 API.
What is an AI agent runtime? (And how it differs from a framework or a chat app)
Frameworks help you write an agent. Chat apps let you talk to one. An agent runtime runs many agents for real users — sessions, tools, sandboxes, memory and channels included.
Frequently asked questions
Everything you need to know about running FastClaw.
Spin up your first agentin under a minute.
Start free in FastClaw Cloud, or install the binary and self-host. Your agents, your infrastructure.
$ curl -fsSL https://raw.githubusercontent.com/fastclaw-ai/fastclaw/main/install.sh | bash