Skip to main content

FastClaw - AI Agent Framework with multi-channel support

Project description

FastClaw

A lightweight but powerful AI Agent framework in Python.

PyPI Installation (Recommended)

# Install
pip install fastclaw-ai

# Run
fastclaw start

# Persistent running (recommended for server)
nohup fastclaw start > /tmp/fastclaw.log 2>&1 &

Configure Agent

After starting, visit http://localhost:8765 , click Settings in the top right corner to configure main_agent, recommended: deepseek-chat.

Or edit the config file:

vim ~/.fastclaw/workspace/data/agents/main_agent/metadata.json

CLI Usage

# Interactive chat
fastclaw chat

# New session
fastclaw chat --new

# Check status
fastclaw status

Git Clone

git clone https://github.com/kandada/fastclaw.git
cd fastclaw
pip install -r requirements.txt
# Run
python main.py start

# Persistent running
nohup python main.py start > fastclaw.log 2>&1 &

Configure Agent

After starting, visit http://localhost:8765 , click Settings in the top right corner to configure main_agent, recommended: deepseek-chat.

Or edit the config file:

vim workspace/data/agents/main_agent/metadata.json

CLI Usage

# Interactive chat
python main.py chat

# New session
python main.py chat --new

# Check status
python main.py status

Design Philosophy

FastClaw is built on the core concept of event-driven agent orchestration. Key design principles:

  • Graph-based Agent: Uses a directed graph to orchestrate agent behavior, supporting conditional branching between nodes (e.g., tool execution, response generation)
  • Event Streaming: Real-time streaming output of LLM responses for better user experience
  • Tool System: Extensible tool framework supporting Shell commands, skills, and custom integrations
  • Session Management: Persistent conversation history with multi-session support
  • Cron Scheduling: Built-in cron-style task scheduling for automated workflows

Features

  • 🤖 LLM-powered - Built on FastMind framework with streaming support
  • 🔧 Tool Calling - Execute Shell commands, skills, and more
  • Cron Jobs - Schedule tasks with cron expressions
  • 💬 Multi-channel - Feishu, iMessage integrations
  • 🎨 Extensible - Easy to add custom skills and agents
  • 🐍 Python Ecosystem - Seamlessly call professional libraries like numpy and pandas, enabling AI to use Python ecosystem tools as skillfully as humans

Changelog

v1.1.8 (2026-05-25)

  • ✨ Feature: LLM Gateway selector (OpenAI / Anthropic) in WebUI agent settings
  • ✨ Feature: Auto-bootstrap default agents from GitHub when workspace is empty
  • ✨ Feature: Auto-generate settings.json with defaults when missing
  • 🔧 Improvement: Increased default Unload Threshold from 80k to 156k

v1.1.7 (2026-05-21)

  • 🐛 Fix: Context unloading now only affects LLM input, messages.jsonl always keeps full history
  • 🐛 Fix: Message timestamps are no longer overwritten on every save
  • 🐛 Fix: CancelledError no longer triggers duplicate response generation
  • 🧪 Test: Added context unloading, sliding window, and timestamp preservation tests

License

GPL-3.0

Copyright (c) 2024-2026 xiefujin. All rights reserved.

Official Website

https://www.fastclaw.world/ | https://fastclaw-ai.com/

Acknowledgments

Inspired by OpenClaw. Special thanks to the open source community.


FastClaw is powerful, but its use depends entirely on the user - all responsibility lies with the user yourself.


Author:xiefujin email: 490021684@qq.com,welcome to contact me.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

fastclaw_ai-1.1.8.tar.gz (628.3 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

fastclaw_ai-1.1.8-py3-none-any.whl (638.8 kB view details)

Uploaded Python 3

File details

Details for the file fastclaw_ai-1.1.8.tar.gz.

File metadata

  • Download URL: fastclaw_ai-1.1.8.tar.gz
  • Upload date:
  • Size: 628.3 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.6

File hashes

Hashes for fastclaw_ai-1.1.8.tar.gz
Algorithm Hash digest
SHA256 fec98bc1a09409cc5b144d6e6a2424ceb074b137eb635b7578f0ea297bd5a584
MD5 535a79f6e491acfa2c547471ea26bf00
BLAKE2b-256 143c00497b1f4ee3ab666d5c0284fa2f37e5dc78c3f074aec044366efcca471f

See more details on using hashes here.

File details

Details for the file fastclaw_ai-1.1.8-py3-none-any.whl.

File metadata

  • Download URL: fastclaw_ai-1.1.8-py3-none-any.whl
  • Upload date:
  • Size: 638.8 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.6

File hashes

Hashes for fastclaw_ai-1.1.8-py3-none-any.whl
Algorithm Hash digest
SHA256 d0208a35880f809e430c7995f3f9afd74b05834a0ec96308dd2f96e15bdd2b08
MD5 7b8b19b9185040a768d37b45bda10267
BLAKE2b-256 e74426e63f570228424f384148206acd4895864edaad6abb3edd50dc6dd4c5e4

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page