Project Orchestrator
A self-hosted AI agent orchestrator that shares code context and task progress across Claude Code, OpenAI Agents, and Cursor. Runs on macOS, Windows, and Linux.
Project Orchestrator is a self-hosted coordination tool for developers working with multiple AI coding agents. It keeps code knowledge, development plans, and architectural decisions in a shared store so agents can reuse context across tasks and projects. It runs on macOS, Windows, and Linux, with desktop apps and Docker deployment available.
Its Rust backend uses Neo4j to record code structure and relationships, while Meilisearch handles searches by meaning. Tree-sitter parses languages including Rust, TypeScript, Python, Go, and Swift. Agents can look up functions, trace imports, and assess which parts of a codebase a change could affect. A file watcher updates the knowledge base as files change.
MCP integration connects that context to Claude Code, OpenAI Agents, and Cursor. Plans track task dependencies and progress, and an autonomous runner can dispatch work to agents in parallel. Related repositories can share a workspace with API contracts and milestones. Notes and decisions link back to the code they concern, so agents can find the reasoning behind an implementation alongside its structure.
Repeatable workflows cover code review and release pipelines, including nested workflows and tracked execution states. Architectural proposals have a review lifecycle that follows them through acceptance and implementation. The backend requires Neo4j and Meilisearch; NATS synchronizes events across instances. The chat interface integrates with Claude, and skills can be exchanged between projects and instances.
