Agentic Council System

A fully free, open-source AI council system built with LangChain and Ollama. This beautiful CLI-based multi-agent system uses 4 specialized AI agents to analyze problems through parallel responses, natural group chat discussions, and weighted synthesis - completely free, no API costs, runs entirely locally.

🆓 100% Free & Open Source

• ✅ No API costs - Runs entirely on your local machine with Ollama
• ✅ No cloud dependencies - All processing happens locally
• ✅ Open source - MIT licensed, fully customizable
• ✅ Privacy-first - Your data never leaves your machine
• ✅ Powered by LangChain - Industry-standard AI framework
• ✅ Powered by Ollama - Free, local LLM inference

🎯 Features

• 🆓 100% Free - No API costs, runs entirely locally with Ollama
• 🎨 Beautiful CLI with rich formatting and colors
• 💬 Natural group chat conversations between agents
• 🏷️ Tag agents with @mentions for focused debates
• ⚖️ Weighted decision model for final recommendations
• 🔍 Comprehensive synthesis analysis
• 🚀 Parallel agent processing for speed
• 📊 Structured output with agreements, conflicts, and blind spots
• 🔒 Privacy-first - All data stays on your machine
• 🛠️ Built with LangChain - Industry-standard AI framework

🏗️ Architecture

The system consists of 4 specialized agents, each with a distinct persona and thinking layer:

1. Elon (Visionary) - First-principles thinking, innovation, bold direction
2. Sam (Strategist) - Business model, market realities, scalability
3. Sheryl (Operator) - Practical execution, system design, reliability
4. Ray (Risk Analyst) - Red-team, failure modes, blind spots

📦 Installation

Prerequisites

1. Python 3.8+ installed

2. Ollama (free, local LLM runtime) - Download from ollama.ai

   # Install Ollama, then start it:
   ollama serve

3. Pull required models (all free):

   # These models run entirely locally - no API costs!
   ollama pull gpt-oss:120b-cloud
   ollama pull glm-4.6:cloud
   ollama pull kimi-k2-thinking:cloud
   ollama pull deepseek-v3.1:671b-cloud

   Note: Model downloads are free. They run locally on your machine.

Install from Source

# Clone the repository
git clone https://github.com/Hrishikeshgupta2002/AI-Council.git
cd AI-Council

# Install dependencies
pip install -r requirements.txt

# Or install as a package
pip install -e .

🚀 Usage

Basic Usage

python main.py

Enter your problem statement when prompted, and the council will analyze it.

Command Line Arguments

# Direct input
python main.py "Your problem statement here"

# Piped input
echo "Your problem statement" | python main.py

Interactive Features

• Normal message: Type your message, all agents respond
• Tag agents: Use @Elon @Sam debate this topic to start a focused debate
• Continue conversation: Press Enter to continue the general conversation
• Exit: Type exit, quit, or q to end

Example

$ python main.py
Enter your problem statement: Should we build this feature?

# Agents respond...

> @Elon @Sam debate the technical approach
# Tagged agents have a focused 2-3 exchange debate

> # Press Enter for another round
# Agents can choose to respond or skip

⚙️ Configuration

Set environment variables to customize behavior:

export OLLAMA_BASE_URL="http://localhost:11434"  # Ollama server URL
export USE_WEIGHTED_MODEL="true"                  # Use weighted model (default: true)
export AGENT_TIMEOUT="60"                         # Agent response timeout in seconds
export MAX_WORKERS="4"                            # Max parallel workers
export DEBUG="true"                                # Show detailed error traces

📁 Project Structure

agentic-council/
├── agents/              # Agent implementations
│   ├── __init__.py
│   ├── visionary.py
│   ├── strategist.py
│   ├── operator.py
│   └── risk_analyst.py
├── council.py           # Council orchestrator
├── synthesis.py         # Synthesis agent
├── state.py             # State management
├── config.py            # Configuration module
├── main.py              # CLI entry point
├── requirements.txt     # Python dependencies
├── setup.py            # Package setup
├── pyproject.toml      # Modern Python packaging
├── LICENSE             # MIT License
└── README.md           # This file

🧩 Module Overview

• council.py: Main orchestrator managing agent interactions
• agents/: Individual agent implementations with distinct personas
• synthesis.py: Meta-agent that synthesizes all responses
• state.py: State management for conversation tracking
• config.py: Centralized configuration management
• main.py: CLI interface with rich formatting

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

💡 Why This is Free

This project is completely free because:

• Ollama provides free, local LLM inference - no API costs
• LangChain is an open-source framework - no licensing fees
• All processing is local - no cloud services or subscriptions needed
• Open source - MIT licensed, you own the code

Unlike commercial AI services that charge per API call, this system runs entirely on your hardware using free, open-source models.

🙏 Acknowledgments

• Built with LangChain - Open-source AI framework
• Powered by Ollama - Free, local LLM inference
• Beautiful CLI with Rich - Terminal formatting library
• All models are free and run locally - no API costs ever!