Routes LLM requests across 300+ models with automatic provider selection based on cost, quality, speed, or balanced scoring. Settles each API call via USDC micropayments on Base L2 using the x402 protocol with sub-cent transaction fees. Exposes six MCP tools for chat completion, session budget management, facilitator health checks, and spending analytics. Install via the embedded VS Code extension or run standalone through npx with Claude Desktop. Useful when you want OpenAI-compatible routing with hard spending caps for autonomous agents or need to optimize inference costs across providers without hardcoding API endpoints. Also implements Google's A2A spec for agent-to-agent communication with payment gates.
claude mcp add --transport stdio io.github.z333q-p402 uvx p402