MassGen is an open-source multi-agent system designed for complex task resolution through collaborative AI. It targets users needing to scale AI capabilities by enabling multiple AI agents to work in parallel, share intelligence, and converge on optimal solutions, inspired by advanced systems like Grok Heavy and Gemini Deep Think.

How It Works

MassGen employs a parallel processing architecture where multiple AI agents, potentially using different models and tools, tackle the same task concurrently. Agents communicate their progress and insights via a shared collaboration hub, allowing for real-time learning and adaptive coordination. The system detects convergence when agents stabilize on a solution through natural collaboration, rather than forced agreement, leading to more robust outcomes.

Quick Start & Requirements

• Installation: git clone https://github.com/Leezekun/MassGen.git, cd MassGen, pip install uv, uv venv.
• API Configuration: Requires API keys for supported models (Anthropic Claude, Google Gemini, OpenAI, xAI Grok) set in a .env file.
• Supported Models: Claude, Gemini, OpenAI, xAI Grok.
• Supported Tools: Web search and code execution are supported depending on the model provider.
• Running: Use uv run python -m massgen.cli --model <model_name> "<your_question>" for single-model tests or uv run python -m massgen.cli --config <config_file.yaml> "<your_question>" for multi-agent configurations.
• Documentation: MassGen README

Highlighted Details

• Supports cross-model/agent synergy by leveraging strengths from diverse frontier models.
• Features parallel processing, intelligence sharing, and consensus building among agents.
• Offers live visualization of agent working processes in real-time.
• Integrates with Claude, Gemini, OpenAI, and Grok models, with potential for local model integration.

Maintenance & Community

• Active development with a roadmap including advanced collaboration, expanded integrations, and a web interface.
• Discord channel available for community support: https://discord.com/invite/VVrT2rQaz5.

Licensing & Compatibility

• Licensed under the Apache License 2.0.
• Permissive license suitable for commercial use and integration with closed-source projects.

Limitations & Caveats

MassGen is in an early stage of development, with potential for upcoming breaking architecture changes. While it supports major LLM providers, integration with local open-weight models is noted as a future enhancement. Grok currently lacks native code execution capabilities.