经 AI Skill Hub 精选评估,智能工作流 获评「推荐使用」。这款Agent工作流在功能完整性、社区活跃度和易用性方面表现出色,AI 评分 7.5 分,适合有一定技术背景的用户使用。
智能工作流 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
智能工作流 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
# 方式一:pip 安装(推荐)
pip install captain-claw
# 方式二:虚拟环境安装(推荐生产环境)
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install captain-claw
# 方式三:从源码安装(获取最新功能)
git clone https://github.com/kstevica/captain-claw
cd captain-claw
pip install -e .
# 验证安装
python -c "import captain_claw; print('安装成功')"
# 命令行使用
captain-claw --help
# 基本用法
captain-claw input_file -o output_file
# Python 代码中调用
import captain_claw
# 示例
result = captain_claw.process("input")
print(result)
# captain-claw 配置文件示例(config.yml) app: name: "captain-claw" debug: false log_level: "INFO" # 运行时指定配置文件 captain-claw --config config.yml # 或通过环境变量配置 export CAPTAIN_CLAW_API_KEY="your-key" export CAPTAIN_CLAW_OUTPUT_DIR="./output"

An open-source AI agent with multi-agent orchestration, autonomous cognitive systems, and a full management dashboard. Runs locally, supports every major LLM provider, and ships with 44 built-in tools.
Glasses Bridge. Captain Claw 0.4.27 introduces a mobile-web → agent → glasses-web pipeline that lets you talk to any Flight Deck agent from your phone and have its reply rendered (and spoken) on Meta Ray-Ban Display smart glasses. Plus Tavily as a web-search provider — thank you to the Tavily team for the excellent API.
captain_claw/flight_deck/glasses_bridge.py) — three pages tied together by a channel-based pub/sub bus in Flight Deck:/glasses/mobile — pick a process agent, type, optionally attach a photo, hit Send. Installable as a PWA on iOS (Share → Add to Home Screen) and Android (install prompt). Standalone mode, dark theme, scoped to /glasses/. Editable channel id in the URL so anyone with the link is in./glasses/view?c=<channel> — floating HUD bar with brand pill + SSR freshness token + live indicator, last 3 messages rendered with full markdown including tables, prominent pulsing "thinking" state when the agent is busy, compact return when the answer arrives./glasses/settings?c=<channel> — tap-target button grids for voice (28) and language (60+), grouped by accent (Neutral / British / American-Spanish / Australian / Indian). Replaces native <select> dropdowns which are essentially unusable with Neural Band gestures.WS /glasses/tts-stream) — first-audio latency 150–300 ms via PCM s16le chunks scheduled through Web Audio API; auto-fallback to one-shot MP3 (POST /glasses/tts) if streaming hiccups. Toggle stream/one-shot in the glasses header (📡 / 📦), persisted across reloads. 60+ languages, 28 multilingual voices.🔇 Voice is off — tap 🔊 to enable banner. Tapping 🔊 doubles as the user gesture that satisfies browser autoplay policy. Zero /glasses/tts* traffic while muted.capture="environment"); bridge proxies the multipart upload to the chosen agent's /api/image/upload and includes the resulting path in the chat. Agent sees the image as the familiar [Attached image: …] prefix.[SYSTEM CONTEXT — do not echo] block telling the model the reply will render on a tiny HUD. Sent to the agent only; never broadcast to the channel bus./glasses/view-manifest.webmanifest so Meta's wearables runtime shows the brand mark instead of the default fallback (Meta does not accept SVG).web_search now supports provider: tavily alongside the default brave. Set TAVILY_API_KEY in env / .env or tools.web_search.tavily_api_key in config. Transparent to the agent — same tool, same query parameter.meta-glasses-test/ — standalone stdlib freshness probe for verifying the "fresh from server every load" behaviour of the glasses webview, separate from the agent stack.Backward compatible — existing 0.4.26 setups keep working unchanged. Glasses Bridge is dormant unless you visit /glasses/*. Tavily is opt-in; the default web-search provider remains Brave. See RELEASE_NOTES_0.4.27.md for the full breakdown and the How to use the Glasses Bridge walkthrough.
See RELEASE_NOTES.md for the full changelog.
docker pull kstevica/captain-claw:latest
docker run -d -p 23080:23080 \
-v $(pwd)/config.yaml:/app/config.yaml:ro \
-v $(pwd)/.env:/app/.env:ro \
-v $(pwd)/docker-data/home-config:/root/.captain-claw \
-v $(pwd)/docker-data/workspace:/data/workspace \
kstevica/captain-claw:latest
See README_DETAILED.md for Docker Compose and persistent data setup.
pip install captain-claw
export OPENAI_API_KEY="sk-..." # or ANTHROPIC_API_KEY, GEMINI_API_KEY, etc.
captain-claw-web # http://127.0.0.1:23080
captain-claw-web # Web UI (default)
captain-claw # Interactive terminal
captain-claw --tui # Terminal UI
captain-claw-fd # Flight Deck multi-agent dashboard
captain-claw-mcp # MCP server for Claude Desktop
botport # Agent-to-agent routing hub
First run starts onboarding automatically. For Ollama, no key needed — set provider: ollama in config.yaml.
YAML-driven with environment variable overrides (CLAW_ prefix).
model:
provider: gemini
model: gemini-2.5-flash
allowed:
- id: claude-sonnet
provider: anthropic
model: claude-sonnet-4-20250514
- id: gpt-4o
provider: openai
model: gpt-4o
web:
enabled: true
port: 23080
Load precedence: ./config.yaml > ~/.captain-claw/config.yaml > env vars > .env > defaults.
Full reference: USAGE.md (23 config sections).
高质量的开源AI工作流项目
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
✅ MIT 协议 — 最宽松的开源协议之一,可自由商用、修改、分发,仅需保留版权声明。
AI Skill Hub 点评:智能工作流 的核心功能完整,质量良好。对于自动化工程师和运维人员来说,这是一个值得纳入个人工具库的选择。建议先在非生产环境试用,再逐步推广。
| 原始名称 | captain-claw |
| 原始描述 | 开源AI工作流:AI agent with multi-agent orchestration, autonomous cognitive systems, and a ful。⭐38 · Python |
| Topics | AIagent工作流 |
| GitHub | https://github.com/kstevica/captain-claw |
| License | MIT |
| 语言 | Python |
收录时间:2026-05-26 · 更新时间:2026-05-30 · License:MIT · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端