repowiserepowise

MCP server

vectifyai / pageindex

πŸ“‘ PageIndex: Document Index for Vectorless, Reasoning-based RAG

β˜… 40,446stars on GitHubPythonMIT

This connects Repowise to vectifyai/pageindex's code. It does not install pageindex.

Repowise maps public repos from their source. Not affiliated with vectifyai. What is Repowise?

github.com/vectifyai/pageindexIndexed at d2693d8

What is pageindex?

PageIndex is a vectorless, reasoning-based RAG engine for long, complex professional documents such as financial reports, legal documents and technical manuals. Instead of splitting documents into chunks and searching vector embeddings, it builds a hierarchical tree index for each document and has an LLM reason its way through that tree to find relevant sections. The project is written in Python and published on PyPI as the pageindex package. To install it, run pip install -U pageindex and follow the steps below. Its code health is 5.1 out of 10 (Needs_work), ranked #239 of 4,000+ indexed repos. The last commit to pageindex was on Sep 24, 2026, and 21 people have contributed.

As of commit d2693d8, indexed Sep 30, 2026. What it is comes from a model reading the README; the rest is measured from the source and git history, not from the maintainer.

For agentspageindex.md

Repo facts: type, license, health, rank, contributors
pageindex facts
TypeMCP server
LicenseMIT
Code health5.1 / 10 (Needs_work)
Health rank#239 of 4,000+ indexed repos
Last commitSep 24, 2026
Contributors21
Bus factor1
Size140 files, 33,728 lines

Install pageindex

Commands copied from the README

  1. 1Install PageIndex from PyPI.
    pip install -U pageindex
  2. 2Import the client class.
    from pageindex import PageIndexClient
  3. 3Submit a PDF to build its tree index and keep the returned document ID.
    doc_id = client.submit_document("report.pdf")["doc_id"]

Releases on GitHub (latest v0.7.4, Nov 19, 2026)

Full install notes on GitHub

Explore pageindex

Ask about this code in your agent

This connects Repowise to pageindex's code. It does not install pageindex. To install pageindex, use the install steps.

Useful if you want to change or fork pageindex: your agent can ask how it is put together and get answers that cite its files.

Ask about pageindex

Answers come from pageindex's README and source, with the files they cite. Two free answers a day without an account.

Want this for your own codebase? Map your repo free

About this page

Get a page like this for your own codebase

Repowise built this map of vectifyai/pageindex from its source and git history. It parses every symbol, draws the dependency graph, scores each file's health from complexity, duplication, coverage and churn, finds hotspots and owners in the history, and writes documentation for every module. Your coding agent reads the same index over MCP, so its answers cite real files.

Free for 2 public repos, with a page like this one. No card. See how Repowise works

Prefer local? pip install repowise && repowise init indexes your own checkout, no account needed. Star Repowise on GitHub

vectifyai/pageindex FAQ

What is PageIndex?

PageIndex is a vectorless, reasoning-based RAG engine that answers questions over long professional documents such as financial reports and legal filings.

How does PageIndex work?

PageIndex builds a hierarchical tree index for each document, then has an LLM search that tree for sections relevant to a question instead of comparing vector embeddings. Each index is saved as JSON and reused across questions.

How do I install PageIndex?

PageIndex is installed from PyPI with pip install -U pageindex. The package then provides PageIndexClient, imported from pageindex. See the install steps

What language is PageIndex written in?

PageIndex is written in Python. Its dependencies include the OpenAI and Anthropic SDKs and the mcp package.

How is the PageIndex code organized?

PageIndex keeps its main code in the pageindex package, with a flash subpackage for PageIndex Flash tree generation. Tests live in a tests directory.

Who is PageIndex for?

PageIndex is for working with long, complex professional documents such as financial reports, legal documents, regulatory filings, technical manuals and academic textbooks.

What license does PageIndex use?

PageIndex is released under the MIT license.

Code health

5.1out of 10Needs work

Hotspots average 3.7, the place fixes pay off most.

On the leaderboards:#68 of 71 Python repos

Key features of pageindex

  • PageIndex retrieves by having an LLM reason over a hierarchical tree index instead of running vector similarity search.
  • PageIndex works without a vector database and without chunking documents.
  • PageIndex returns results that are traceable to explicit references in the source document.
  • PageIndex runs locally on your machine with your own LLM key, or against PageIndex Cloud with an API key.
  • PageIndex Flash generates tree indexes for text-based PDFs and is the default indexing method in local mode.

How vectifyai/pageindex works

Documents enter as files such as PDFs and are passed to PageIndexClient, which is imported from pageindex. pageindex/flash/main.py is one of the two entry points and runs the PageIndex Flash pipeline that builds the tree index for text-based PDFs. The pageindex/flash/clustering directory groups visual and positional evidence into stable regions and proposes candidates that later assembly can use as structural anchors. The pageindex/flash/labels directory models blocks, lines, spans and caption-like areas as bounded regions, so captions are handled as regions rather than stray text. pageindex/local_chat.py is one of the hotspot files for chat in local mode, where an LLM answers questions from the retrieved parts of the tree, and each finished index is saved as a JSON file that later questions reuse.

The codebase has 140 files and 2,040 symbols in 4 modules, led by Python, JavaScript and Shell. Code health is 5.1 out of 10, rated needs work.