MADE: a Rust engine for structured multi-agent deliberations
MADE lets teams of AI agents run formal, ceremony‑driven deliberations with peer critique, validation, and LLM‑as‑judge scoring. It enforces output contracts via JSON Schema, records decisions as replayable OpenTelemetry traces, and exposes a contract‑first gRPC API plus an MCP stdio server. Designed for developers building reliable AI workflows, it offers observability and Prometheus metrics that make debugging and auditability straightforward, standing out from generic orchestration tools.
View on GitHub →underpass-ai/made