EtienneChollet/ontomics
View on GitHub ↗Extract domain knowledge from codebases to reduce LLM token consumption by 20x and time in agentic search by 10x — gathers and makes concepts, naming conventions, and vocabulary queryable via MCP.
40 ★8 forksRustUpdated 4mo ago
What you need to know
Codebase instant-knowledge MCP server that builds an ontology of a codebase so agents can answer questions with fewer tokens. Implemented in Python, Rust and TypeScript.
Key features
- ✓Instant codebase knowledge via ontology
- ✓Fewer tokens than raw code context
- ✓Python, Rust and TypeScript implementations
Best for
Efficient codebase Q&A in agents with limited context
Clients: MCP-compatible clients
Documentation ↗Reviewed 2026-08-11
Topics
agentic-codingai-codingclaude-codecodexcoding-agentsdomain-knowledgemcpmcp-servermodel-context-protocolrustsemantic-searchtree-sitter
- Stars
- 40★
- Forks
- 8
- Language
- Rust
- License
- MIT
- Created
- 2026-03-27
- Last push
- 2026-04-17