Skip to content
WhichAgent

LlamaIndex

Agent framework by LlamaIndex, Inc.

ActiveOpen source · MITVerified Oct 22, 2026
Website Docs Source

Open-source Python framework for building LLM agents and retrieval (RAG) pipelines over your own data, from the company behind the LlamaParse document-parsing platform.

Key facts

Type
Agent framework · Library
Pricing
Free
Free tier
The LlamaIndex framework is free and MIT-licensed. The hosted LlamaParse platform (formerly LlamaCloud) has a free plan with 10K credits/month.
License
MIT
Models
OpenAI, Anthropic, Google Gemini, AWS Bedrock, Azure OpenAI, Mistral, Local models via Ollama, vLLM, llama.cpp, 100+ via LLM integration packages incl. LiteLLM
Languages
Python
Platforms
macOS, Windows, Linux
Bring your own API key
Yes
Local models
Yes
MCP support
Yes
Self-hostable
Yes
GitHub stars
52,617
Last commit
Oct 14, 2026

Best for

  • RAG and question-answering over private documents
  • Agents that need many data connectors and vector store integrations
  • Teams already using LlamaParse for document parsing

Pros

  • Deep tooling for ingestion, indexing, and retrieval
  • Large catalog of LLM, embedding, vector store, and data-loader integrations
  • Event-driven Workflows for multi-step agents
  • MCP tools can be used and LlamaIndex tools/workflows exposed via MCP

Cons

  • TypeScript version is deprecated, so it is effectively Python-only
  • Still pre-1.0 (0.14.x), with periodic breaking changes
  • Many features are split across separately versioned integration packages

How LlamaIndex compares

LangChain vs LlamaIndex

For agents, pick LangChain.

Updated Oct 22, 2026

Sources (9)