vectifyai / pageindex
📑 PageIndex: Document Index for Vectorless, Reasoning-based RAG
github.com/vectifyai/pageindexIndexed at d2693d8Up to date with upstream
About this codebase
PageIndex is a vectorless, reasoning-based RAG system that consumes document files (including PDFs), builds a reasoning-oriented document tree with the PageIndex Flash pipeline, and then produces retrieval-augmented answers with citations via SDK local mode or PageIndex Cloud; it outputs a persisted per-document JSON index plus query-time chat responses (and tree retrieval results) that an agent can call through MCP or a direct API surface. In practice, developers use the PageIndex SDK to `index` (tree-build), then `retrieve` (tree search over the stored structure), and finally `chat` (LLM responses grounded in the retrieved nodes), reusing the same index artifacts across many questions.
Written by gpt-5.4-nano from the index
140 files and 2,040 symbols in 4 modules, led by Python (109 files) and JavaScript (1).
Start reading
- Knowledge graph140 files and how they depend on each other.
- Char-Level PDF Content WalkingDocumentationDocumentation for pageindex/flash/parser_pdfium_charlevel, written by gpt-5.4-nano.
- Introduce local mode for PageIndex SDK with Flash tree build and OpenAI-style chatDecisionProposed decision, from a pull request.
Code health
This codebase scores 5.1 out of 10 for code health, which we rate needs work. It also scores maintainability 5.6 and static performance 9.7 out of 10. The three are scored separately and never blended into one number. Risk is concentrated, as it usually is: 19 of 140 files are git hotspots, and they average 3.7 — which is where the fixes pay off most.
Do next
1 thing worth doing in the week to Sep 29, the last indexed commit, 1 of them now.
Now
- Critical
Clean up what this week's commits left: 5 serious findings in 4 files
1 commit in the last 7 days added or worsened them, and they are still open. Start with
pageindex/flash/api.py, which has 1 critical.- serious findings
- 5
- critical
- 2
- files
- 4
Medium effort
Improve what Repowise can see
Add a test coverage report
Without one, Repowise cannot tell tested code from untested code, so every test-related action says “unknown”.
- line coverage
- Unknown, no report
Small effort
Use vectifyai/pageindex in your agent
Point Claude Code, Cursor, Codex or VS Code at this repository. Your agent gets the index this page is built from: cited answers, callers, history and health for any file.
https://api.repowise.dev/mcp/vectifyai/pageindexOne command, run anywhere. Adds the server to your local Claude Code config.
Docsclaude mcp add --transport http repowise https://api.repowise.dev/mcp/vectifyai/pageindex \ --header "Authorization: Bearer rw_live_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
This repo is public, so anyone with a Repowise account and an API key can connect.
Try asking
- get_overview()
- get_answer("how is pageindex put together?")
- get_why("why is pageindex structured this way?")
Prefer local? pip install repowise && repowise init indexes your own checkout, no account needed.
Go deeper
- OverviewEverything the index found, what changed this week, commits and decisions.
- DocumentationModule-by-module documentation, 246 pages, 15 written by a model.
- ArchitectureDependency graph, layers and third-party dependencies.
- Code healthEvery file's score, the findings behind it, tests and refactoring plans.
- CommitsEach commit's effect on code health, with agent provenance.
- ContributorsWho owns what, and where one person holds most of the knowledge.
- DecisionsArchitectural decisions mined from commits and pull requests.
- Ask this codebaseSign inGrounded answers over the indexed code.
For maintainers
Add a badge to your README
Both badges are public, cached, and update on their own after every index. Nothing to install.
Repowise badge
Links to this page. Adding it also re-indexes the repo every week.
[](https://repowise.dev/repo/vectifyai/pageindex)Code health badge
Average health across every file, straight from the latest index.
[](https://repowise.dev/repo/vectifyai/pageindex)How vectifyai/pageindex works
📑 PageIndex: Document Index for Vectorless, Reasoning-based RAG This page is a map of the vectifyai/pageindex repository, written primarily in Python, rebuilt from the source each time it is indexed. Repowise parses every symbol, computes a dependency graph, scores per-file code health from complexity, duplication, test coverage and churn, mines git history for hotspots and ownership, and lifts the architectural decisions into documentation you can read here or query through MCP.
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.
Use the links above to open each view, or connect this repository to your agent for grounded answers inside Claude, Cursor, Codex or VS Code.