LlamaIndex

LlamaIndex is a document-agent stack: an MIT Python/TS framework plus hosted LlamaParse. Rechecked 2026-08-17: llamaindex.ai/pricing sells Free $0 (10K credits), Starter $50 (40K credits, pay-as-you-go up to 400K), Pro $500 (400k credits, optional one-time 800k / $1,000 bonus on business-email domains, usage up to $5,000 / month), and Enterprise custom. The page says 1,000 credits = $1.25. github.com/run-llama/llama_index had 51,696 stars, license MIT. The GitHub blurb now leads with document agents and OCR, not a generic "LLM application framework" essay.

Compare Dify if you wanted a visual RAG canvas with public $59 / $159 seats, LangChain if you wanted another MIT agent library, and RAGFlow if you wanted an Apache-2.0 RAG appliance.

Key Features

• OSS framework: Connectors, indexes, query/chat engines. Treat current docs as the catalog.
• LlamaParse credits: Hosted parsing/extraction is what the public dollar table meters. Basic parse can be 1 credit / page; agentic/VLM parse costs more.
• Cloud vs library: You can stay on GitHub and pay nothing to LlamaIndex. Credits start when you use LlamaParse / cloud actions.
• Not a free OCR unlimited: Free is 10K credits.

Use Cases

• Teams parsing messy PDFs who will pay Starter $50 or Pro $500.
• Teams that only need the MIT library and will stay on Free / self-host.
• People comparing LlamaIndex vs Dify vs RAGFlow on public meters.

Limitation: We did not re-open every connector list. Startup-program free credits exist; do not invent a dollar amount.

Pricing

llamaindex.ai/pricing on 2026-08-17. This is the LlamaParse / platform table.

If a leftover blog still describes LlamaIndex as an unpriced framework only, treat this table as the source for hosted parse.

Getting Started

1. Open llamaindex.ai/pricing before you parse a 10k-page corpus.
2. Use the MIT library from developers.llamaindex.ai if you do not need LlamaParse.
3. Recheck Dify or RAGFlow if you wanted a visual workspace.
4. Quote 1,000 credits = $1.25 when you do the napkin math.

Frequently Asked Questions

Is LlamaIndex free?

The library is MIT. Hosted parse starts at Free 10K credits, then $50 / $500.

Same as LangChain?

Both are MIT agent/RAG libraries. LlamaIndex's public dollar table is LlamaParse credits.

Same as Dify?

No. Dify is a visual cloud workspace. LlamaIndex is a library plus a parse meter.

Alternatives

• Dify: Visual RAG/agent cloud.
• LangChain: MIT agent library.
• RAGFlow: Apache-2.0 RAG engine.

Tips

1. Quote Free 10K / Starter $50 / Pro $500.
2. Do not paste "unpriced LLM framework" into a deck.
3. Recheck agentic parse credit burn on scanned PDFs.

Conclusion

LlamaIndex is a MIT framework with a public LlamaParse credit ladder, not a generic connector slogan. Start at llamaindex.ai/pricing, then decide whether Dify already covers the workspace you need.