Gradient Bang offers an online multiplayer universe where players explore, trade, battle, and collaborate, with a unique integration of LLMs as active participants. It targets players seeking dynamic, AI-driven gameplay and developers interested in real-time agentic workflows, providing a platform to experience advanced multi-tasking, tool-calling, and low-latency voice interactions within a persistent world.

How It Works

The project's core innovation lies in its fully agentic design, where every entity, including player ships, is an AI agent. It leverages a real-time architecture built on Supabase for its PostgreSQL database and Edge Functions, serving as the game server. A React client handles the UI, while Pipecat Cloud hosts the AI agents. This approach enables complex, emergent behaviors and interactions between human players and AI agents, showcasing advanced capabilities like multi-tasking and sophisticated tool use.

Quick Start & Requirements

For users with Claude Code installed, the fastest setup is the /init command, which automates dependency installation, Supabase setup, environment file generation, world data creation, and API key prompting. Otherwise, manual setup requires:

• Prerequisites: uv (Python package manager), Supabase account, Docker, Node.js 18+. Optional: Pipecat Cloud account, Supabase CLI.
• Setup: Involves starting Supabase locally via Docker, configuring environment variables (.env.supabase, .env.bot, .env.cloud), running edge functions, starting the Pipecat bot, and launching the React client.
• Links: Claude Code skills reference is available within the project's documentation.

Highlighted Details

• Demonstrates advanced real-time agentic workflows, including multi-tasking and tool calling.
• Features low-latency voice integration for agent communication.
• Generates a universe map visualization (artifacts/universe-map.svg) using scripts/universe_svg.
• Offers a local pooler mode to reduce Supabase Edge Function latency by running logic directly against Postgres.

Maintenance & Community

The provided README does not detail specific community channels (e.g., Discord, Slack), notable contributors, or a public roadmap. The project integrates with Claude Code, offering automated workflows via slash commands.

Licensing & Compatibility

The README does not specify a software license. Therefore, compatibility for commercial use or closed-source linking cannot be determined without further information.

Limitations & Caveats

Setting up Gradient Bang locally involves managing multiple distributed components (Supabase, edge functions, client, bot), which can be complex. The system requires numerous API keys for various services (e.g., Deepgram, Google AI, Anthropic/OpenAI, Supabase, Pipecat Cloud), posing an adoption barrier. While aiming for low-latency voice, performance may vary, and specific configurations are needed for optimal local development.