An agent factory: many agents, each with its own soul
Create and run many AI agents from one FastClaw install — each with its own SOUL.md personality, memory, skills, model and sharing settings.
Personality as files
SOUL.md, IDENTITY.md, AGENTS.md, TOOLS.md and friends define who an agent is. Edit them in the console or push them from the CLI.
Long-term memory
MEMORY.md holds durable facts and is refreshed by the heartbeat. Full session history is preserved, thinking content included.
Per-agent overrides
Give each agent its own model, temperature, sandbox setting, private skills and IM channels — without touching the others.
Public sharing
Flip Public access and anyone with the chat link can talk to the agent under their own account, with their own sessions and memory.
Export and import
Export an agent’s identity files, prompt mode and own skills as a ZIP, and import one to replace them after a preview. Keys and chats stay out.
CLI and dashboard, one store
fastclaw agents init creates a chattable agent in one command. Agents made in the CLI show up in the web UI and vice versa.
What an agent is made of
FastClaw is an Agent Factory: it creates, manages and runs agents, and handles the LLM calls, tool execution, sandboxing and sessions underneath. Each agent is a set of system files — SOUL.md for personality, IDENTITY.md, USER.md, MEMORY.md, HEARTBEAT.md, AGENTS.md, TOOLS.md, BOOTSTRAP.md and agent.json — plus its own model choice, skills and channels.
Open an agent in the console and you get a management panel: Chat for testing, Files for the system files, Skills, Models, Channels, Scheduler and Sessions.
From zero to a chattable agent
fastclaw agents init alpha --provider openai --model openai/gpt-4o-mini --api-key-env OPENAI_API_KEY
fastclaw agents files put alpha SOUL.md ./SOUL.md
The CLI writes through the same store the dashboard uses and signals the running gateway to hot-reload. Need a starting point for the personality? The SOUL.md generator drafts one for you.
Many agents, many people
Agents can be private or shared. A public agent keeps one identity while every chatter gets their own history and memory. Admins can provision agents for other users and fork an existing agent's identity into a new account — the building block for "user buys a bot" flows that agencies and SaaS builders rely on. For a personal setup, see FastClaw for developers.
Works with
All Integrations
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.
Telegram
Turn any FastClaw agent into a Telegram bot in minutes: paste a BotFather token, pair your account, and chat in DMs or groups.
Frequently asked questions
More Features
All FeaturesModels
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.
Try it on your own agents
Start free in FastClaw Cloud, or self-host the single binary in a minute.