Repolode
Submit
run-llama

run-llama/llama_index

A Python framework for building RAG, agent and document-processing applications that connect LLMs to your own data.

View on GitHubWebsite
Stars
52,578
Forks
8,387
Open issues
997
Last commit
12 hours ago

What is it?

LlamaIndex is an open-source Python framework for building LLM applications on top of your own data. It provides data connectors that ingest files, APIs and databases, indexing structures such as vector stores and graphs that organize that data for retrieval, and query interfaces that feed retrieved context into an LLM to produce grounded answers.

The core package, llama-index-core, is extended through a large set of integration packages covering LLM providers, embedding models and vector stores. Developers can start with the batteries-included llama-index package or install only the integrations they need. High-level APIs let beginners build a vector index and query it in a few lines, while lower-level components such as retrievers, query engines and rerankers can be customized.

The project also includes agent and workflow tooling for multi-step applications. Its maintainers have shifted their commercial focus toward LlamaParse, a hosted document parsing and extraction platform that can be used alongside or independently of this framework.

Key features

  • 01Data connectors for ingesting files, APIs, databases and other sources.
  • 02Vector store, graph and other index structures for organizing data for retrieval.
  • 03Query engines and retrievers that pass retrieved context to an LLM.
  • 04Modular core with over 300 integration packages for LLMs, embeddings and vector stores.
  • 05Agent and workflow tools for multi-step applications.
  • 06Index persistence to disk for reloading without rebuilding.

Use cases

  • Building a question-answering chatbot over internal company documents.
  • Prototyping a RAG pipeline that swaps between different LLM and vector store providers.
  • Creating agents that call tools and reason over private data sources.
  • Indexing PDFs, web pages or databases for semantic search with an LLM.

Quick start

  1. 1

    Install the starter package

    pip install llama-index
  2. 2

    Build a vector index over your documents with an OpenAI key set

  3. 3

    Query the index

    query_engine = index.as_query_engine()

Things to know

  • The README says the company now focuses mainly on LlamaParse, so the open-source framework gets less attention.
  • Default examples call paid hosted LLM APIs such as OpenAI unless you configure a local model like Ollama.

Topics

#rag#llm#agents#vector-database#python#data-framework#application#data#fine-tuning#framework

FAQ

What is llama_index?
A Python framework for building RAG, agent and document-processing applications that connect LLMs to your own data. LlamaIndex is an open-source Python framework for building LLM applications on top of your own data. It provides data connectors that ingest files, APIs and databases, indexing structures such as vector stores and graphs that organize that data for retrieval, and query interfaces that feed retrieved context into an LLM to produce grounded answers.
Is llama_index free to use? What license does it have?
llama_index is open source under the MIT license. Check the license terms before using it in commercial products.
Is llama_index still maintained?
Actively maintained. The last commit to llama_index was on Oct 20, 2026.
What can I use llama_index for?
Building a question-answering chatbot over internal company documents. Prototyping a RAG pipeline that swaps between different LLM and vector store providers. Creating agents that call tools and reason over private data sources. Indexing PDFs, web pages or databases for semantic search with an LLM.
How do I install llama_index?
Install the starter package: pip install llama-index Build a vector index over your documents with an OpenAI key set Query the index: query_engine = index.as_query_engine()
Where can I get llama_index?
The source code and installation instructions are on GitHub: https://github.com/run-llama/llama_index

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects