Show HN: Nia – MCP server that gives more docs and repos to coding agents
Hacker News (score: 22)Description
Coding agents generate code well but lose accuracy when the answer lives outside the repo in front of them. Developers end up pasting GitHub links, docs, and blog posts by hand and hoping the agent scrolls far enough. Long context windows help, but recent “context rot” measurements show quality still drops as prompts grow. For example, in LongMemEval, all models scored much higher on focused (short, relevant) prompts (~300 tokens) than on full (irrelevant, 113k tokens) prompts, with performance gaps persisting even in the latest models (https://research.trychroma.com/context-rot).
Nia is a MCP that gives more context to any coding agent or IDE. It Indexes multiple repos and docs sites and makes this available via MCP to your coding agent so it has much more context to work with, giving you more specific and accurate answers.
Nia uses a hybrid code search architecture that combines graph-based structural reasoning with vector-based understanding. When a repo or documentation is ingested, Tree-sitter parses it into ASTs across 50+ languages and natural languages, and the code is chunked by function/class boundaries into stable, content-addressable units. These chunks are stored both in a graph db to model relationships like function calls and class inheritance, and in a vector store. At query time, a lightweight agent with give_weight tool dynamically assigns weights between graph and vector search based on intent (e.g., "who calls X" vs "how does auth work"), and both paths are searched in parallel. Results are fused, enriched with full code context, and passed through multi-stage rerankers: semantic reranker, cross-encoders, LLM-based validators.
Early Signal: In internal evals we improved Cursor’s performance by 27 % once Nia had indexed external docs models couldn’t get from their training data or searching the web.
Quickstart: <https://www.youtube.com/watch?v=5019k3Bi8Wo> Demo: <https://www.youtube.com/watch?v=Y-cLJ4N-GDQ>
To try it out: grab an API key at https://app.trynia.ai/ and follow instructions at https://docs.trynia.ai/integrations/nia-mcp.
Try it and break it! I’d love to know which contexts your agent still misses. Corner cases, latency issues, scaling bugs. I’m here 24/7.
Thanks!
More from Hacker
Show HN: Self-updating MCP server for official pip, uv, poetry and conda docs
Show HN: Self-updating MCP server for official pip, uv, poetry and conda docs
HNSW as abstract data structure: video intro to Redis vector sets
HNSW as abstract data structure: video intro to Redis vector sets
Show HN: I rewrote an outdated React Native map clustering library
Show HN: I rewrote an outdated React Native map clustering library Hey Hacker News,<p>I'm a long-time lurker and wanted to share a project I just finished building.<p>Like many React Native developers, I needed to add marker clustering to a map in my app. The most popular library for this, react-native-maps-clustering, was fantastic in its day but has become outdated and no longer works with modern versions of Expo, React Native, and their dependencies.<p>After hitting a wall of compatibility issues, I decided to take on the challenge of rewriting it from the ground up, focusing on a modern toolchain and a better developer experience.<p>The journey was a lot more challenging than I anticipated. It turned into a deep dive into solving dependency hell with different versions of @types/react, wrestling with build tool configurations for pnpm, bob, and ESLint, and ensuring everything was strictly typed with TypeScript. It felt like a classic case of yak shaving, but I was determined to create a solution that "just works" for developers today.<p>The result is RN Super Cluster, a performant, fully-typed, and easy-to-use clustering library for react-native-maps.<p>What it does: It provides a <ClusteredMapView /> component that you can use as a drop-in replacement for the standard <MapView />. Any <Marker /> components you place inside will be automatically clustered.<p>Key Features:<p><pre><code> Modern & Maintained: Built with a modern toolchain and designed to be actively maintained. Fully-Typed: Written entirely in TypeScript to prevent common errors and improve autocompletion. High-Performance: Uses supercluster under the hood for extremely fast geospatial clustering. Spiderfier: At the maximum zoom level, overlapping markers automatically "spiderfy" (spread out on a spiral) so they can be individually tapped. Customizable: You can provide your own custom components for rendering clusters, and callbacks for handling press events. </code></pre> This was a passion project born out of necessity, and I hope it can save other React Native developers the headaches I went through.<p>I would love to get your feedback, and contributions are more than welcome!<p>GitHub: <a href="https://github.com/suwi-lanji/rn-maps-clustering">https://github.com/suwi-lanji/rn-maps-clustering</a> NPM: <a href="https://www.npmjs.com/package/rn-maps-clustering" rel="nofollow">https://www.npmjs.com/package/rn-maps-clustering</a><p>Thanks for checking it out!
No other tools from this source yet.