AI Tools · AI Development

PageIndex

PageIndex is Vectify AI's vectorless RAG engine that indexes long PDFs as a tree for cited answers, available open source or as a cloud API.

Our take: PageIndex is one of the more talked-about RAG approaches this year, and the repository's popularity backs that up.

Best forLong-document RAG

ReviewedOct 2026Freemium

7.55 criteria · see breakdown

Try PageIndexFree (MIT); Cloud $0.01/page after $10 credit

PageIndex is a vectorless RAG engine from Vectify AI: it builds a tree index of long documents such as financial reports or contracts and lets an LLM reason over it to return cited answers. The open-source version is free under the MIT licence; PageIndex Cloud starts with $10 in free credits, then charges $0.01 per indexed page.

The bottom line

A strong option for cited Q&A over long PDFs, if you accept LLM-heavy retrieval.

Our score

Reviewed by the Conversion Gems editorial team. Hover a facet to focus it; click to pin.Tap a facet to focus it.

What it is, and why we rate it

One line each. Open a row for detail.

What it really isAn open-source and cloud document index that lets an LLM reason over a tree of sections instead of searching vector embeddings.

Our takePageIndex is one of the more talked-about RAG approaches this year, and the repository's popularity backs that up.

For long, well-structured documents where you must show exactly where an answer came from, the tree index and citations make sense. Keep two things in mind: every query still runs through an LLM you pay for, and the headline accuracy numbers are Vectify AI's own. Test it on your documents before you commit.

Why we rate itA free MIT core, clear per-page Cloud pricing and traceable citations for long documents.

The catchLLM costs on every query, vendor-run benchmarks, and text-only PDFs in the local version.

Best / not good for

Best for

Q&A over financial reports and filings
Contract and manual search with citations
Teams avoiding a vector database

Not good for

Fast-changing data that needs constant re-indexing
Teams without an LLM budget
Scanned PDFs without using Cloud

Trade-offs

Friction report

Trade-offs to know before signing up.

Time to value
Minutes on Cloud: install the SDK, add a PageIndex API key and an LLM key, then submit a PDF.
Scale breakpoint
Large collections that change often, since each re-indexed page is billed on Cloud and every query uses LLM tokens.
Walled garden
Low: MIT-licensed SDK, bring-your-own LLM, and an Enterprise option in your own cloud.

Still a fit?

If none of these block you, start on the free plan.

Try PageIndex

PageIndex pricing

Option	Price	What you get
PageIndex Local (open source)	$0	MIT-licensed SDK on your machine, text-based PDFs, your own LLM key
Cloud free start	$10 in free credits	No payment method; covers about the first 1,000 pages indexed
Cloud indexing	$0.01 per page	Charged once, when a page is indexed
Cloud active pages	$0.001 per page per month	First 1,000 active pages free each month
Cloud retrieval	No PageIndex fee	Unlimited queries on active pages; your LLM provider bills you
Enterprise	Quote	Runs in your VPC on AWS, Azure or GCP, with audit logs, dedicated support and SLA

From PageIndex's pricing docs and developer page, checked 4 October 2026. Subscribers on the older Standard, Pro or Max plans keep them until they cancel.

Frequently Asked Questions

What is PageIndex?

PageIndex is a vectorless, reasoning-based retrieval (RAG) engine made by Vectify AI. It builds a tree index of a long document and lets an LLM work through that tree to find the sections that answer a question, instead of matching vector embeddings.

Is PageIndex free?

The open-source version is free under the MIT licence: you run it locally and pay only your own LLM provider. PageIndex Cloud starts with $10 in free credits and no payment method, then charges by usage (checked 4 October 2026).

How much does PageIndex Cloud cost?

Indexing costs $0.01 per page, charged once. Pages kept active cost $0.001 per page per month, with the first 1,000 free. Retrieval on active pages has no PageIndex fee; your LLM provider bills you directly. The docs' own example: 3,000 pages cost $30 to index, then $2 a month (checked 4 October 2026). Enterprise is quote-based.

Is PageIndex on GitHub?

Yes. The open-source repository is VectifyAI/PageIndex, MIT-licensed, with about 38.6K stars when we checked. You install it with pip install -U pageindex. The quickstart uses an OpenAI API key, and the developer page says any model works.

What does vectorless RAG mean?

Retrieval without embeddings or a vector database. PageIndex keeps the document's structure as a tree and asks an LLM to pick the relevant branches. Vectify AI says this avoids the chunking and similarity-search errors of vector RAG; test it on your own documents to confirm.

What is the difference between PageIndex Local and Cloud?

Local runs on your machine with your own model key, keeps documents on your disk and reads text-based PDFs without OCR. Cloud adds managed indexing and storage, scanned and image-rich files, block-level citations and an MCP server.

PageIndex alternatives

Other tools for document retrieval and RAG:

Showing 1–3 of 4

Sources checked October 4, 2026: PageIndex developer page, PageIndex pricing docs, PageIndex on GitHub, PageIndex Chat, About Vectify AI.

Conversion Gems independently reviews every tool.

PageIndex

PageIndex pricing

Frequently Asked Questions

PageIndex alternatives

Tags

Explore related categories