Meinrag — a multimodal RAG platform

Document intelligence for PDFs that fight back: layout-aware parsing, hybrid retrieval with reranking, and an MCP layer so agents can use it as a tool.

PythonFastAPIPostgreSQLFAISSDoclingMCPDocker

GitHub

The problem

Real-world document corpora are hostile to naive RAG: scanned PDFs, dense tables, figures that carry the actual answer, mixed languages, thousands of files. Chunk- and-embed alone retrieves the wrong paragraph with great confidence.

What I built

A self-hosted RAG platform that treats retrieval quality as a pipeline, not a model call:

What it proves

Retrieval is an engineering discipline: parsing fidelity, retrieval staging, and honest evaluation move the needle more than model choice. The failure modes I hit — index/registry desync, dedup keyed on the wrong hash, per-article recall diluted by fixed-size chunking — each became a design rule, and several were found by measuring recall on real questions rather than eyeballing demos.

Status

In production since spring 2026, queried daily — including by the agent fleet, which uses it through the MCP layer.