English | 简体中文

🍊 Akka - Self-learning, 24/7 Creator Agent

Complete the full loop from topic selection to publishing with a single sentence

---

📖 Table of Contents

• 🍊 Akka - Self-learning, 24/7 Creator Agent
  • 🎬 Demo Video
  • 🎯 Core Capabilities
  • 🚀 Quick Start
  • ✨ Features
  • 📖 User Guide
  • 📂 Project Structure
  • 🔌 API Endpoints
  • 🙏 Acknowledgments
  • 📧 Contact Us
  • 📄 License

---

🎬 Demo Video

Click to watch the demo video

---

🎯 Core Capabilities

Akka (Automated Knowledge & Kontent Assistant) is an open-source self-learning, 24/7 Creator Agent that achieves a complete loop from topic selection to publishing through browser automation, material search, content generation, and content publishing.

Just one instruction, Akka automatically completes:

Capability	Description
🤖 Automated Browser Operations	Search competitors, collect materials, view data
🔍 Automated Material Search	Intelligently search and organize relevant materials
🎨 Automated Content Generation	Outline planning, copywriting creation, image generation
📤 Automated Publishing	Automatically publish content

From topic planning to scheduled publishing, let AI Agent boost your social media operation efficiency by 10x, truly achieving "what you think is what you get".

🚀 Quick Start

3-Step Launch

# 1. Clone the project
git clone https://github.com/living-stream/akka.git
cd akka

# 2. Configure API Key
cp config.example.yaml config/config.yaml
# Edit config/config.yaml and fill in your OpenAI API Key

# 3. Start
./start.sh

Visit http://localhost:3000 to get started!

Prerequisites

• Python 3.10+
• Node.js 18+
• Chrome Browser

Configuration

Option 1: Environment Variables (Recommended)

export OPENAI_API_KEY="your-api-key"
export OPENAI_BASE_URL="https://api.openai.com/v1"  # optional

Option 2: Configuration File

llm:
  provider: "openai"
  api_key: "your-api-key"
  model: "gpt-4o"

✨ Features

🖥️ Web Interface

Feature	Description
💬 Smart Chat	Streaming conversation experience with thinking process display
📝 Workspace	Note card display, image thumbnail preview, click to expand details
⏰ Scheduled Tasks	Create and manage scheduled tasks with real-time status tracking
⚙️ Settings	User persona and memory configuration

🤖 Intelligent Content Generation

• Fully Automated Creation: Automatically generate complete Xiaohongshu (Little Red Book) image-text notes from a single instruction
• Three-stage Generation Flow: Outline planning → Copywriting creation → Image generation
• Image Generation: Automatically generate supporting images, supporting multiple images

📊 Skill System

Skill	Function
competitor-benchmarking	Competitor benchmarking analysis
content-creation	Content generation (including browser information collection)
content-review	Content review analysis
inspiration-hunting	Inspiration collection
scheduled-task	Scheduled task management

🌐 Browser Automation

• Information Collection: Search competitors, collect store information, view account data
• Platform Operations: Support for Xiaohongshu, Dianping and other platforms

🧠 Intelligent Control

• User Isolation: Each user has independent soul.md, memory.md, workspace
• Conversation Memory: Automatically save conversation history, auto-compress when exceeding 20 entries
• Task Lock: Ensure only one task executes at a time for the same user

📖 User Guide

Web Interface

1. Chat Page - Chat with AI assistant, give task instructions
2. Workspace - View generated note content, click cards to expand details
3. Tasks Page - Manage scheduled tasks, view task status
4. Settings Page - Configure user persona and memory

Command Line Client

Execute Task:

python master/client.py --uid test_user run "Help me analyze competitors in the coffee sector"

More Examples:

# Create scheduled task
python master/client.py --uid test_user run "Help me create a task to publish a coffee post tomorrow at 12pm"

# Generate note
python master/client.py --uid test_user run "Write a note about pour-over coffee"

# Competitor analysis
python master/client.py --uid test_user run "Search for viral notes in the coffee sector on Xiaohongshu, analyze their titles and cover characteristics"

Health Check:

python master/client.py health

📂 Project Structure

akka/
├── master/                    # Backend core
│   ├── server.py              # FastAPI server
│   ├── client.py              # Command line client
│   ├── controller.py          # CoreController core logic
│   ├── scheduler.py           # Scheduled task scheduler
│   └── skills/
│       └── definitions/       # Skill definitions directory
├── web/                       # Next.js frontend
│   ├── app/                   # Page routes
│   │   ├── chat/              # Chat page
│   │   ├── workspace/         # Workspace page
│   │   ├── tasks/             # Tasks page
│   │   └── settings/          # Settings page
│   ├── components/            # UI components
│   │   ├── ui/                # Basic components
│   │   ├── chat/              # Chat components
│   │   ├── workspace/         # Workspace components
│   │   └── tasks/             # Task components
│   ├── hooks/                 # React Hooks
│   ├── store/                 # Zustand state management
│   └── lib/                   # Utility functions
├── users/                     # User data directory
│   └── {uid}/                 # User-specific directory
│       ├── soul.md            # User persona
│       ├── memory.md          # Long-term memory
│       ├── conversation.json  # Conversation history
│       ├── tasks.json         # Scheduled tasks
│       └── workspace/         # Workspace directory
│           └── {note_folder}/
│               ├── copywriting.md  # Note copywriting
│               ├── outline.md      # Outline
│               └── images/         # Images
├── agentic_tool/              # Agent tools
│   ├── browser_use_agent/     # Browser automation
│   └── note_generate_agent/   # Content generation
├── llm/                       # LLM model factory
├── config/                    # Configuration files
└── start.sh                   # One-click startup script

🔌 API Endpoints

Conversation

Method	Path	Description
POST	/run	Execute task (streaming response)
GET	/conversation/{uid}	Get conversation history
DELETE	/conversation/{uid}	Clear conversation history

Workspace

Method	Path	Description
GET	/workspace/{uid}	Get note list
GET	/workspace/{uid}/notes/{path}	Get note content
GET	/workspace/{uid}/images/{folder}/{name}	Get note images

Scheduled Tasks

Method	Path	Description
POST	/schedule	Create scheduled task
GET	/tasks/{uid}	Get task list
DELETE	/tasks/{uid}/{task_id}	Cancel task

User

Method	Path	Description
GET	/user/{uid}/profile	Get user profile
PUT	/user/{uid}/soul	Update user persona
PUT	/user/{uid}/memory	Update user memory

🙏 Acknowledgments

Akka is built on the following excellent open-source projects:

• browser-use - Browser automation framework
• LangChain - LLM application development framework
• LangGraph - Multi-agent orchestration framework
• FastAPI - Modern Python web framework
• Next.js - React full-stack framework

Thanks to the developers and communities behind these projects!

📧 Contact Us

If you have any questions or suggestions, feel free to contact us:

• Email: burns@ven-ai.top

📄 License

MIT License