knomi¶
knomi is a token-efficient document ingestion CLI and RAG connector for local AI agents: it indexes your documents into a vector database (Qdrant or ChromaDB) and serves them as a retrieval source for Claude, OpenWebUI, Ollama and other agents.
- Install:
pip install knomi(PyPI) - Source: github.com/franjofranjic27/knomi
- Usage: see the README
Where to start¶
| Page | Content |
|---|---|
| Architecture | Ingestion pipeline, interfaces, serve mode |
| ADR-001 Vector Store | Why Qdrant default + pluggable stores |
| ADR-002 Token Chunking | Why chunk sizes are measured in tokens |
| ADR-003 SHA-256 Dedup | How unchanged documents are skipped |
| Commit Convention | Commit message format |
| Testing | Test types and how to run them |
| CI/CD Workflows | GitHub Actions pipelines, releases |
| Troubleshooting | Common problems and fixes |
For contributor setup see CONTRIBUTING.md; repo-wide standards live in franjofranjic27/.github.