0.5.4

AgentCrew MCN 🤖

📖 中文文档 | English

---

What is AgentCrew?

AgentCrew MCN is an open-source multi-agent content marketing automation tool. A team of AI "employees" handles the entire pipeline — from content creation to cross-platform distribution.

• Writer Agent — Copywriter: generates technical articles, social posts, and threads
• Reviewer Agent — QA: pre-publish content safety and quality review
• Publisher Agent — Operations: publishes to Juejin, Zhihu, Dev.to, and more
• Analyst Agent — Data Analyst: tracks performance and smart scheduling

Features

• 🧠 4 AI Agents — Writer, Reviewer, Publisher, Analyst — each operates independently
• 🔧 Skills + Tools System — Atomic operations composed into high-level capabilities (v0.4: LLM-driven)
• 📚 RAG Knowledge Base — Retrieval-augmented generation from historical content
• 🌍 Cross-Platform Publishing — Juejin, Zhihu, Dev.to with platform-specific style adaptation
• 🔌 Multi-Provider LLM — DeepSeek, OpenAI, Anthropic, Ollama — switch with one config field
• 🖼️ Image Generation — DALL-E 3 cover images for articles
• ⏰ Smart Scheduling — Analyst predicts best publish times with random jitter anti-detection
• 📊 Dashboard + Growth Tracking — Web UI for analytics + GROWTH.md for metrics
• 🐳 Docker — One-command deployment with docker-compose
• 🔁 Dogfooding — The project promotes itself

Quick Start

Option 1: pip install (recommended)

# 1. Install
pip install agentcrew-mcn

# 2. Initialize config
agentcrew-mcn init

# 3. Edit .env with your API keys
#    DEEPSEEK_API_KEY=sk-...

# 4. Start using
agentcrew-mcn write generate --topic "Python async programming" --style technical
agentcrew-mcn publish post --file article.md --platform juejin --dry-run

💡 agent-crew is also available as an alias for agentcrew-mcn.

Option 2: From source

# 1. Clone
git clone https://github.com/super-rick/agentcrew-mcn.git
cd agentcrew-mcn

# 2. Install dependencies
pip install -r requirements.txt

# 3. Initialize
python -m cli.main init

# 4. Edit .env with your API keys

Configuration

agentcrew-mcn init creates a config template. Key settings:

# config.yaml
llm:
  api_key: ${DEEPSEEK_API_KEY}  # Reads from .env
  model: deepseek-chat

platforms:
  juejin:
    cookie: ${JUEJIN_COOKIE}    # Export cookie from browser
  devto:
    api_key: ${DEVTO_API_KEY}   # https://dev.to/settings/extensions

Usage

# Generate a technical article
agentcrew-mcn write generate --topic "Python async programming" --style technical

# Preview mode (no LLM call, inspect parameters)
agentcrew-mcn write generate --topic "Python async" --dry-run

# Publish to Juejin
agentcrew-mcn publish post --file article.md --platform juejin

# Scheduled publishing (every 6 hours)
agentcrew-mcn schedule start --topic-file topics.txt --platform juejin --interval 6

# RAG knowledge base
agentcrew-mcn rag ingest --file article.md --source "my_blog"
agentcrew-mcn rag search --query "AI Agent architecture"

Architecture

┌─────────────────────────────────────────────────────┐
│                   Orchestrator                       │
│           Task dispatch / Scheduler / Config          │
└──┬──────────┬──────────────┬────────────────────────┘
   │          │              │
┌──▼──────┐ ┌▼────────┐ ┌──▼────────┐
│ Writer  │ │Publisher│ │  Analyst  │
│ Agent   │ │Agent    │ │  Agent    │
│         │ │         │ │           │
│ Skills: │ │ Skills: │ │ Skills:   │
│ trending│ │ schedule│ │ report    │
│ tech    │ │ multi   │ │ recommend │
│ thread  │ │ batch   │ │           │
│         │ │         │ │           │
│ Tools:  │ │ Tools:  │ │ Tools:    │
│ search  │ │ devto   │ │ analytics │
│ rag     │ │ juejin  │ │ compare   │
│ compose │ │ zhihu   │ │ report    │
└─────────┘ └─────────┘ └───────────┘

Dashboard

AgentCrew includes a Streamlit Dashboard for monitoring your AI marketing team.

More Dashboard Pages

Publish Analysis	AI Analysis

Tech Stack

Roadmap

See ROADMAP.md for details.

• v0.1 ✅ Writer + Publisher + CLI + Juejin/Zhihu + Dashboard
• v0.2 ✅ Dev.to adapter + RAG + MCP + Bilingual README
• v0.3 ✅ Reviewer Agent + Retry + Persistence + Docker (332 tests)
• v0.4 ✅ LLM-driven Skills + Multi-provider + Image gen + Smart scheduling
• v0.5 ✅ 6 new platforms (CSDN/WeChat/SegmentFault/X/Xiaohongshu/Medium)
• v1.0 🔮 Web API + Plugins + Community

License

MIT