Data framework for connecting LLM agents to private and structured knowledge sources.
LlamaIndex is a data framework focused on connecting LLM agents to private, structured, or semi-structured knowledge sources — documents, databases, and APIs — via indexing and retrieval pipelines. It's for developers building retrieval-augmented generation (RAG) systems and agents that need grounded, up-to-date data access. It differs from general orchestration frameworks by specializing deeply in data ingestion, indexing, and retrieval rather than broad agent-executor tooling.
| Tier | Price | Notes |
|---|---|---|
| Open source | $0 | MIT licensed, self-hosted |
| LlamaCloud | $0–$250/mo | Free tier plus paid managed parsing/indexing tiers |
| Enterprise | Contact us | Custom volume, SSO |
Yes, the core framework is open source and free under the MIT license. LlamaCloud's managed parsing/indexing has a free tier and paid tiers up to $250/month.
LangChain covers similar retrieval use cases within a broader orchestration framework; LlamaIndex specializes more narrowly in data indexing.
Yes, LlamaIndex supports Model Context Protocol servers as both tool sources and data connectors.
RAG and data retrieval are its core focus, but it also supports agent workflows built on top of its indexing and query layers.
No, it ships both Python and TypeScript SDKs.