frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: Docdex – A local tool to reduce LLM tokens and make agents smarter

https://github.com/bekirdag/docdex
1•bekirdag•1h ago
Hi HN,

I use LLMs every day for software development, and wanted to have a better experience by reducing the token usage and providing digested information about the codebase to the agent.

So I built Docdex. The idea was simple: a local, persistent layer that preprocesses and structures your project so the model can spend its context window on the actual problem, not on rediscovering what already exists.

It started as a document indexer in Rust, built on Tantivy for proper ranked full-text search.

Then it kept growing. I added code search, AST-based symbol indexing, a directed symbol graph for impact analysis ("what breaks if I change this function?"), repo memory for project-level facts, and agent memory so the LLM can persist preferences across sessions.

One architectural decision I'm particularly happy with: it runs as a single daemon. MCP servers have a habit of getting stuck, each tool call spawning processes, LLMs not cleaning up after themselves, multiple SSE connections fighting over resources. Running everything as one lightweight background service sidesteps most of that. It handles multiple repos, re-indexes on changes, and stays out of the way.

I also added web search, but handled locally. During setup it can install Ollama, pull `nomic-embed-text` for embeddings, and optionally `phi3.5` for lightweight reasoning. Search results get scored and filtered locally before anything hits a paid API. Brave Search works best, but Google/Bing work too with API keys.

There's also job delegation, expensive frontier models can offload simpler subtasks to a local Ollama model, which cuts paid token usage noticeably in practice.

Supported languages for impact graph reasoning: Rust, Python, JS/TS, Go, Java, C++, and more.

You can also use it standalone, as a local search engine for your docs, a plain HTTP server, or a web search and summarization tool. It's not glued to any particular AI workflow.

How to install: npm i -g docdex

Setup: docdex setup

Website: https://docdex.org, Repo: https://github.com/bekirdag/docdex

Fully open source, free to use. API keys for web search are optional.

It's genuinely made my daily dev workflow better, fewer tokens, more consistent answers, less time reexplaining things. I plan to keep developing it and may eventually offer optional paid services (like delegated cloud model jobs via OpenRouter), but the core will stay open.

Would love feedback, especially from people building serious AI dev tooling.

Show HN: Merge – Connect Apple Watch to Android

https://www.merge.watch/apple-watches/
1•urip•20s ago•0 comments

Show HN: Aguara – Security scanner for AI agent skills and MCP servers

https://github.com/garagon/aguara
1•garagon•26s ago•0 comments

California importing oil from Bahamas as state faces highest US gas prices

https://nypost.com/2026/02/16/us-news/desperate-california-now-shipping-oil-from-the-bahamas-usin...
1•mhb•1m ago•0 comments

Pmarchive – An archive of the best articles from Marc Andreessen's 2007 blog

https://pmarchive.com/
1•hisamafahri•1m ago•0 comments

Show HN: Ambulatory glucose profile analysis tool

https://github.com/daedalus/agp
1•dclavijo•1m ago•0 comments

Show HN: I built a website for generating viral shorts with AI

https://aiclipsfarm.com/
1•lastFitStanding•2m ago•0 comments

The Token Production System

https://pavpanchekha.com/blog/tps.html
1•pavpanchekha•3m ago•0 comments

Show HN: API Combat – A game played through API calls

https://apicombat.com/
1•mcarthey•4m ago•0 comments

We Used CRDTs to Build Real-Time Collaborative Retrospectives

https://rootly.com/blog/how-we-used-crdts-to-build-real-time-collaborative-retrospectives
1•sylvainkalache•4m ago•0 comments

OK, It's a Bubble. Now Tell Me How It Pops

https://fullhoffman.com/2026/02/14/ok-its-a-bubble-now-tell-me-how-it-pops/
1•mooreds•4m ago•0 comments

Show HN: Beyblade Restock Alerts on Amazon

https://www.pricedropnotifications.com/beyblade-restock-alerts
1•hntime•6m ago•0 comments

Show HN: Tropes.fyi – Name and shame AI writing

https://tropes.fyi/
1•ossa-ma•7m ago•0 comments

Star Trek Removed from Amazon Prime [TNG and Voyager Refunded]

https://oldcynic.com/star-trek-removed-from-amazon-prime-tng-voyager-refunded
2•xvxvx•8m ago•1 comments

The Prisoner's Dilemma of Super Bowl Ad Economics

https://philippdubach.com/posts/the-economics-of-a-super-bowl-ad/
1•7777777phil•8m ago•0 comments

Destroy My Startup

https://shipordie.club/roast/startup
1•dsego•10m ago•0 comments

Show HN: OSS Durable Memory for LLMs

1•Mnexium•11m ago•0 comments

Polymaths

https://www.kakeiho.com/polymaths/
1•erhuve•12m ago•0 comments

Linking Smaller Haskell Binaries

https://brandon.si/code/linking-smaller-haskell-binaries/
1•todsacerdoti•13m ago•0 comments

Amazon service was taken down by AI coding bot

https://www.ft.com/content/00c282de-ed14-4acd-a948-bc8d6bdb339d
2•ceejayoz•14m ago•1 comments

The iOS Weekly Brief – Issue #48

https://iosweeklybrief.com/p/the-ios-weekly-brief-issue-48
1•khambir•15m ago•0 comments

Variable interpolatable smooth curves and outlines

https://thottingal.in/blog/2026/02/20/var-interpolatable-smooth-curves/
1•sthottingal•15m ago•0 comments

Gen AI: Time to Focus on Logic Quality and Not Code Quality

https://balanarayan.com/2026/02/20/gen-ai-time-to-focus-on-logic-quality-and-not-code-quality/
1•speckx•16m ago•0 comments

She Wanted a Quiet House in the Berkshires for About $800k

https://www.nytimes.com/interactive/2026/02/19/realestate/berkshire-mountains-massachusetts-homes...
1•whack•16m ago•0 comments

Child's Play: Tech's new generation and the end of thinking

https://harpers.org/archive/2026/03/childs-play-sam-kriss-ai-startup-roy-lee/
2•ramimac•17m ago•0 comments

Show HN: Natural language search across Kalshi and Polymarket

https://www.attena.xyz/
1•helloiamvu•17m ago•0 comments

Ente Locker

https://ente.io/locker/
2•amanverasia•19m ago•0 comments

Power generation increased by up to 70% when wind were combined with tidal/wave

https://www.surrey.ac.uk/news/power-generation-increased-70-cent-when-wind-turbines-were-combined...
2•JeanKage•20m ago•0 comments

Session-graph – A knowledge graph from your AI coding sessions

https://github.com/robertoshimizu/session-graph
1•rshimizubr•20m ago•1 comments

Where Lisp Fails: At Turning People into Fungible Cogs (2009)

https://www.loper-os.org/?p=69
3•tosh•20m ago•1 comments

Show HN: AstrMap – Unix Philosophy for the AI Era (Ditch the RAG)

https://github.com/hubby247/astrmap
1•hubby247•21m ago•0 comments