AI Potluck
Product / UX / Orchestration & agents

LlamaIndex

LlamaIndex

Data framework in Python and TypeScript for building LLM applications over private data. It provides connectors across APIs, PDFs, documents and SQL, indices, retrievers, query engines and reranking, and an agent and workflows orchestration layer. Its depth is in retrieval and ingestion - document agents, many-format parsing and structured extraction.

LlamaCloud and LlamaParse are sold beside the library as document-processing services and gate nothing in it. Verified 2026-08-11 via the run-llama/llama_index repository tree and the vendor pricing page.

Openness

5 high confidence
5.0
license
MIT(core)
commercial
LlamaCloud/LlamaParse
source
public
core-gated
ungated(every package in run-llama/llama_index is MIT with no enterprise directory

MIT self-hostable core with a commercial managed cloud beside it. A full untruncated recursive tree of run-llama/llama_index at main is 13,820 entries whose license files are the root MIT plus a per-package copy in every llama-index-integrations package and in llama-index-core, with no enterprise, ee, commercial or proprietary path anywhere, and llamaindex.ai/pricing prices LlamaParse alone in document-processing credits, listing LlamaIndex, Workflows and LiteParse separately under "Open Source". Nothing is withheld from the published library: that the vendor has a paid product at all says nothing about the core, and LlamaParse is a different product sold beside it rather than a piece cut out of it.

  • https://github.com/run-llama/llama_index recorded 2026-06-17

    MIT, ~50.2k stars, LlamaCloud commercial tier

  • https://api.github.com/repos/run-llama/llama_index/git/trees/main?recursive=1 recorded 2026-08-11

    Full untruncated recursive tree, 13,820 entries. License files are the root LICENSE, llama-dev/LICENSE, llama-index-core/LICENSE and a per-package LICENSE in each llama-index-integrations/* package. No enterprise/, ee/, commercial/ or proprietary/ path anywhere.

  • https://api.github.com/repos/run-llama/llama_index recorded 2026-08-11

    Repo metadata: language Python, license spdx_id MIT, size 482,901 KB, 51,564 stars, default branch main.

  • https://www.llamaindex.ai/pricing recorded 2026-08-11

    Titled 'LlamaParse Pricing: Compare Plans & Credits'. Prices only LlamaParse document-processing credits: Free $0 / 10K credits, Starter $50 / 40K, Pro $500 / 400K, Enterprise custom with SSO and hybrid deployment. The nav lists LlamaIndex, Workflows and LiteParse separately under 'Open Source - OSS repos trusted by millions of developers'. No paid tier of the llama_index library.

Adoption

4 high confidence
4.0

7,203,303 PyPI downloads in the trailing 30 days for llama-index, which bands at level 4 (1M-10M) on the software and model adoption scale. The figure briefly cleared 10M and has since fallen back.

Capability

5 high confidence
5.0

Best-in-class RAG/data-ingestion depth plus an agent layer.

  • https://github.com/run-llama/llama_index recorded 2026-08-13

    README - data connectors across APIs, PDFs, docs and SQL; indices, retrievers, query engines and reranking modules; document agents built with Workflows and Agent Builder; 51.6k stars, 7.9k forks

Verified 2026-08-11