indx

pip install "indx[cloud]"
export OPENAI_API_KEY="sk-..."
indx ./docs --out ./ai-ready

The file to directory thesis

A parser answers “what does this file say?” An agent asks a harder question: “where does this file belong, what does it depend on, and what surrounding context should I trust?” A folder is not a bag of files. It has shape, lineage, and implied relationships. indx keeps that map and turns it into something agents and RAG systems can use.

Directory-level

The unit of work is the directory, not the file. Nested trees, ZIPs, and mixed formats flow through one pipeline into one coherent knowledge space.

Relationship-aware

Folder hierarchy, sibling files, and cross-document references become a typed graph. An agent learns that /contracts/2024/ means something.

Semantic metadata

Document type, topics, tags, and summaries are attached as metadata, so retrieval can filter and reason instead of guessing.

Portable output

The result is a self-contained, versioned .indx archive with a readable index.json, per-chunk files, and portable embeddings. Build it once, ship it anywhere.

Bring your own stack

Parser, LLM, VLM, embedder, vector store, output: every slot is a typed interface with a sensible default. No lock-in.

Local-first

Use --offline for the zero-dependency core path, or install indx[local] for local parsing, enrichment, embeddings, and storage.

Start here

What is indx? The problem, the thesis, and how indx composes parsers instead of replacing them.

Installation Install the light core and the optional extras that power each component slot.

Quickstart From pip install to a queryable knowledge space in 60 seconds.

Tutorial Build, inspect, and query your first knowledge space end to end.

Go deeper

Core Concepts KnowledgeSpace, Document, Chunk, Relation — the small mental model the whole API rests on.

The Pipeline Walk → Parse → Chunk → Relate → Enrich → Embed+Pack, one replaceable stage at a time.

Guides Configure components, run air-gapped, write custom stages, author plugins, and tune performance.

Reference CLI, SDK, data models, protocols, the .indx format, and the index.json schema.