Open-source Jev + PageIndex searches 300-page docs without a vector DB or embeddings
gaganghotra_ · x · 2026-10-03
VectifyAI open-sourced jev-doc-search, showing how to build long-document search (300-page reports) with Jev's Choice calls plus PageIndex — no vector database, no embeddings.
Key idea: frame "which page answers this question?" as a multiple-choice problem where Jev returns a probability per option. When the document exceeds what Jev can read at once, PageIndex turns it into a hierarchical tree, and Jev descends node by node, choosing among a handful of options at each level — scaling to arbitrary document lengths.
The repo includes both flat page search and tree search implementations.
More from coding & agent
- Sign in with ChatGPT lets Plus/Pro users spend plan usage across 11 partner apps — _AustinCalvert_ · 2026-10-03
- Best tech stacks for agentic coding: training data volume decides everything — aiblastoff · 2026-10-03
- After auditing dozens of companies: the 4 reasons custom AI workflows fail in production — salespire · 2026-10-03
- No Moat in AI Agents: Zero-Cost Switching Leaves Only User Setup as a Barrier — _AustinCalvert_ · 2026-10-03
- Cloudflare Open-Sources Its Internal AI 'OS', Hits 10k GitHub Stars — ritakozlov · 2026-10-03
- Dev Builds Modular MOSS Rover That Runs on Pi 5, Jetson and Any Depth Camera — OpenAIDevs · 2026-10-03