MIMIRQ DOCUMENTATION

Make knowledgepaths explainable.

Start with a working deployment, then inspect parsing, governance, chunking, retrieval, reranking, citations, and regression gates one layer at a time.

Start here Read the full guide

KNOWLEDGE PATH

AssessParseGovernChunkIndexRetrieveRerankCiteEvaluate

Choose the shortest path to your goal

The handbook is organized by tasks, not by feature inventory.

First deployment

From .env to first login

Prepare credentials, choose Docker or host processes, and verify readiness.

Deploy MimirQ

First knowledge base

From dataset to cited answer

Upload a document, inspect chunks, test retrieval, and trace the final evidence.

Follow the workflow

Existing application

Integrate API, Dify, or a workflow

Use stable contracts for auth, ingestion, retrieval, streaming, and retries.

Open integration guide

SYSTEM MAP

Inspect the system by layer

Each area documents its input, output, boundaries, failure modes, and verification path.

Knowledge operationsDatasets, ingestion, parsing, chunking, retrieval, and evidence.Integration flowsEnd-to-end scenarios, tenant headers, SSE, retries, and Dify.OperationsDeployment, configuration, health probes, and observability.

The code and the handbook evolve together.

OpenAPI remains the schema source of truth. The handbook explains how those contracts form a reliable knowledge workflow.

Browse API reference GitHub