{"slug": "chroma", "name": "Chroma", "category": "databases", "category_name": "Databases", "description": "", "recommenders": 6, "commenters": 23, "warners": 0, "score": 9.517, "rising": -0.64, "first": "2025-11-21", "latest": "2026-09-16", "why": "I'm not colorblind but I use that quite often to check roughly if the color palette I choose is fine", "why_is_quote": true, "has_verdict": false, "sentiment": {"recommend": 6, "neutral": 17, "compare": 1, "warn": 0}, "spark": [0, 2, 1, 0, 0, 1, 0, 0, 1, 0, 1, 0], "listed": true, "loved": [], "caveat": "", "short_quote": "I'm not colorblind but I use that quite often to check roughly if the color palette I choose is fine", "kind": "", "icon": true, "domain": "trychroma.com", "github_repo": null, "official_url": "https://www.trychroma.com", "pricing": "", "aliases": ["Chroma", "ChromaDB"], "verdict": null, "praise": [{"label": "", "summary": "", "quotes": [{"id": 795545073, "item_id": 49721567, "url": "https://news.ycombinator.com/item?id=49721567", "by": "ToriTech", "date": "2026-09-16", "version": null, "stance": "recommend", "quote": "I use ChromaDB with Python", "excerpt": "Assuming you want a plan with a $0 budget, it's probably best if you start with TryHackMe and Portswigger Academy. The free tier on TryHackMe gives you access to a lot of labs and challenges, mostly focusing on Web security, forensics, and some blue team work, but they're starting to add more AI content as of late. Portswigger Academy is completely free, made by the people who created Burpsuite, and has a lot of high", "context": "", "use_case": "vector database for prompt injection experiments", "switched_from": "", "story": "Started Learning Pentesting"}, {"id": 783602848, "item_id": 48975178, "url": "https://news.ycombinator.com/item?id=48975178", "by": "nmdra", "date": "2026-07-19", "version": null, "stance": "recommend", "quote": "ingesting note chunks into ChromaDB while preserving all Wikilinks and tags", "excerpt": "Hi, Huge fan of your work! Just to clarify, this project isn't an alternative to obsidian-cli. My goal is simply to enable semantic search across my vault by ingesting note chunks into ChromaDB while preserving all Wikilinks and tags. This allows AI agents to retrieve specific, semantically relevant text chunks based on user queries, rather than fetching entire notes. It also helps to find hidden connections between ", "context": "", "use_case": "semantic search over notes", "switched_from": "", "story": "Show HN: Local-first CLI to make Obsidian vaults searchable for AI agents"}, {"id": 766100320, "item_id": 47881270, "url": "https://news.ycombinator.com/item?id=47881270", "by": "Narew", "date": "2026-04-23", "version": null, "stance": "recommend", "quote": "I'm not colorblind but I use that quite often to check roughly if the color palette I choose is fine", "excerpt": "Ubisoft release some tools to simulate colorblindness. It's easy to extract de color transformation from their shader directly. I'm not colorblind but I use that quite often to check roughly if the color palette I choose is fine https://github.com/ubisoft/Chroma", "context": "not colorblind, designs palettes", "use_case": "colorblindness simulation", "switched_from": "", "story": "Your hex editor should color-code bytes"}, {"id": 743815809, "item_id": 46488488, "url": "https://news.ycombinator.com/item?id=46488488", "by": "brabel", "date": "2026-01-04", "version": null, "stance": "recommend", "quote": "Used this for code syntax higlighting: https://github.com/alecthomas/chroma", "excerpt": "I did the same thing, but implemented my site generator in Go. My site has grown by a lot over the years, but I can still build it from scratch (from MD files, HTML snippets and static files) in less than one second! Also have a RSS feed generator and it can highlight code in most programming languages, which is important to me as I write posts on many languages. I did try Hugo before I went on to implement my own, a", "context": "", "use_case": "code syntax highlighting", "switched_from": "", "story": "Jeffgeerling.com has been Migrated to Hugo"}, {"id": 740590035, "item_id": 46286877, "url": "https://news.ycombinator.com/item?id=46286877", "by": "gardnr", "date": "2025-12-16", "version": null, "stance": "recommend", "quote": "I use Chroma for a lot of projects.", "excerpt": "pgVector is great and so is FAISS, but those are just a subset of what you get from Milvus. If all you need to do is RAG over 50Mb of documents then pick the right tool for the job. I use Chroma for a lot of projects. Then, what if you want hybrid search, or different IVF variants, or disk-based search, or horizontal scaling, or something that leverages SIMD, or sparse vectors? Milvus is great.", "context": "", "use_case": "", "switched_from": "", "story": "7 Years, 2 Rebuilds, 40K+ Stars: Milvus Recap and Roadmap"}]}], "caveats": [], "who": [{"by": "rasikjain", "context": "Generative AI Engineer", "url": "https://news.ycombinator.com/item?id=48369910", "date": "2026-06-02"}, {"by": "Narew", "context": "not colorblind, designs palettes", "url": "https://news.ycombinator.com/item?id=47881270", "date": "2026-04-23"}, {"by": "philip1209", "context": "runs a Mac Mini home datacenter", "url": "https://news.ycombinator.com/item?id=46635273", "date": "2026-01-15"}], "tags": [], "use_cases": ["vector storage", "vector database for prompt injection experiments", "semantic search over notes", "vector search", "colorblindness simulation", "vector store for knowledge base"], "compared": [{"slug": "qwen", "name": "Qwen", "co": 4, "switched_from": 0, "switched_to": 0}, {"slug": "react", "name": "React", "co": 4, "switched_from": 0, "switched_to": 0}, {"slug": "fastapi", "name": "FastAPI", "co": 4, "switched_from": 0, "switched_to": 0}, {"slug": "qdrant", "name": "Qdrant", "co": 4, "switched_from": 0, "switched_to": 0}, {"slug": "gpt-4", "name": "GPT-4", "co": 3, "switched_from": 0, "switched_to": 0}, {"slug": "langgraph", "name": "LangGraph", "co": 3, "switched_from": 0, "switched_to": 0}], "makers": [{"id": 760536321, "item_id": 47533520, "url": "https://news.ycombinator.com/item?id=47533520", "by": "philip1209", "date": "2026-03-26", "version": null, "stance": "neutral", "quote": "I built Chroma's integration for Stripe Projects.", "excerpt": "I built Chroma's integration for Stripe Projects. As a developer tool, integrating Stripe Projects felt a lot like adding \"Sign in with Google\" - Stripe acts as a trusted identity and billing provider, but for agents instead of humans. The core insight is that agent commerce is a trust problem: an agent can't (shouldn't?) enter a credit card or verify an email, so you need a trusted third party to KYC both sides. Str", "context": "", "use_case": "", "switched_from": "", "story": "Stripe Projects: Provision and manage services from the CLI"}, {"id": 755590432, "item_id": 47224402, "url": "https://news.ycombinator.com/item?id=47224402", "by": "philip1209", "date": "2026-03-02", "version": null, "stance": "neutral", "quote": "Chroma is the open-source search database for AI applications.", "excerpt": "Chroma | ONSITE - San Francisco | Full-time | https://trychroma.com Chroma is the open-source search database for AI applications. It supports dense vector, sparse vector, regex, and full-text search. We support local, single-node, and cloud hosting. We're a team of 14, all engineers today. We work in Rust, Python, Typescript, and Go. We're hiring for these areas: - Account executive (founding) - Database storage eng", "context": "", "use_case": "", "switched_from": "", "story": "Ask HN: Who is hiring? (March 2026)"}, {"id": 749750000, "item_id": 46859375, "url": "https://news.ycombinator.com/item?id=46859375", "by": "philip1209", "date": "2026-02-02", "version": null, "stance": "recommend", "quote": "Chroma is the open-source search database for AI applications.", "excerpt": "Chroma | ONSITE - San Francisco | Full-time | https://trychroma.com Chroma is the open-source search database for AI applications. It supports dense vector, sparse vector, regex, and full-text search. We support local, single-node, and cloud hosting. We're a team of 14, all engineers today. We work in Rust, Python, Typescript, and Go. We're hiring for these areas: - Account executive (founding) - Database storage eng", "context": "", "use_case": "AI search database", "switched_from": "", "story": "Ask HN: Who is hiring? (February 2026)"}, {"id": 743489648, "item_id": 46468103, "url": "https://news.ycombinator.com/item?id=46468103", "by": "philip1209", "date": "2026-01-02", "version": null, "stance": "recommend", "quote": "Chroma is the open-source search and retrieval database for AI applications.", "excerpt": "Chroma | ONSITE - San Francisco | Full-time | https://trychroma.com Chroma is the open-source search and retrieval database for AI applications. It supports dense vector, sparse vector, regex, and full-text search. We support local, single-node, and cloud hosting. We're a team of 14, all engineers today. We work in Rust, Python, Typescript, and Go. We're hiring for these areas: - Product design - Production engineeri", "context": "on a 14-engineer team hiring for Chroma", "use_case": "Vector search database", "switched_from": "", "story": "Ask HN: Who is hiring? (January 2026)"}, {"id": 737324288, "item_id": 46082768, "url": "https://news.ycombinator.com/item?id=46082768", "by": "0xC45", "date": "2025-11-28", "version": null, "stance": "recommend", "quote": "I would also recommend checking out Chroma", "excerpt": "For an open source, local (or cloud) vector DB, I would also recommend checking out Chroma (https://trychroma.com). It also supports full text search. Disclaimer: I work on Chroma cloud.", "context": "", "use_case": "Open-source vector database", "switched_from": "", "story": "So you wanna build a local RAG?"}, {"id": 732838400, "item_id": 45802400, "url": "https://news.ycombinator.com/item?id=45802400", "by": "kbdiaz", "date": "2025-11-03", "version": null, "stance": "recommend", "quote": "Chroma is the open-source database designed for AI.", "excerpt": "Chroma | ONSITE - San Francisco | Full-time | https://trychroma.com Chroma is the open-source database designed for AI. Help build the future of context engineering! We're a team of 17, all engineers today. We work in Rust, Python, Typescript, and Go. We're hiring for these areas: - Database Storage - Distributed Systems - Product Engineering - Platform - Product Design What we\u2019ve recently done: - We built and launch", "context": "", "use_case": "Open-source AI database", "switched_from": "", "story": "Ask HN: Who is hiring? (November 2025)"}], "threads": [{"url": "https://news.ycombinator.com/item?id=46616529", "title": "Ask HN: How are you doing RAG locally?", "mentions": 3}, {"url": "https://news.ycombinator.com/item?id=47161620", "title": "Show HN: Director-AI \u2013 token-level NLI+RAG", "mentions": 2}, {"url": "https://news.ycombinator.com/item?id=49711533", "title": "Started Learning Pentesting", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49522896", "title": "Ask HN: Who wants to be hired? (September 2026)", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48974142", "title": "Show HN: Local-first CLI to make Obsidian vaults searchable for AI agents", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48747975", "title": "Ask HN: Who wants to be hired? (July 2026)", "mentions": 1}], "sentiment_by_month": {"2025-11": {"neutral": 4}, "2025-12": {"recommend": 2, "neutral": 3}, "2026-01": {"recommend": 1, "compare": 1, "neutral": 2}, "2026-02": {"neutral": 1}, "2026-03": {"neutral": 3}, "2026-04": {"recommend": 1}, "2026-05": {"neutral": 1}, "2026-06": {"neutral": 1}, "2026-07": {"neutral": 1, "recommend": 1}, "2026-08": {}, "2026-09": {"neutral": 1, "recommend": 1}, "2026-10": {}}, "versions": [], "version_unnamed": 0, "family_ref": null, "runs_on": null, "lines": [], "quotes": [{"id": 795545073, "item_id": 49721567, "url": "https://news.ycombinator.com/item?id=49721567", "by": "ToriTech", "date": "2026-09-16", "version": null, "stance": "recommend", "quote": "I use ChromaDB with Python", "excerpt": "Assuming you want a plan with a $0 budget, it's probably best if you start with TryHackMe and Portswigger Academy. The free tier on TryHackMe gives you access to a lot of labs and challenges, mostly focusing on Web security, forensics, and some blue team work, but they're starting to add more AI content as of late. Portswigger Academy is completely free, made by the people who created Burpsuite, and has a lot of high", "context": "", "use_case": "vector database for prompt injection experiments", "switched_from": "", "story": "Started Learning Pentesting"}, {"id": 792390486, "item_id": 49524405, "url": "https://news.ycombinator.com/item?id=49524405", "by": "sjtroxel", "date": "2026-09-01", "version": null, "stance": "neutral", "quote": "Chroma, Pinecone", "excerpt": "Location: Cape Girardeau, MO, USA (CST) Remote: Yes Willing to relocate: No Technologies: Python, FastAPI, TypeScript, React, LangGraph, RAG, Chroma, Pinecone, pgvector, PostgreSQL, Redis Streams, Anthropic/OpenAI/Gemini APIs, MCP, LLM-as-judge evals, AWS Lambda + Bedrock, Docker, GitHub Actions R\u00e9sum\u00e9/CV: https://sjtroxel.github.io Email: sjtroxel@protonmail.com --- I'm basically self-taught / career-changer. Everyt", "context": "", "use_case": "", "switched_from": "", "story": "Ask HN: Who wants to be hired? (September 2026)"}, {"id": 783602848, "item_id": 48975178, "url": "https://news.ycombinator.com/item?id=48975178", "by": "nmdra", "date": "2026-07-19", "version": null, "stance": "recommend", "quote": "ingesting note chunks into ChromaDB while preserving all Wikilinks and tags", "excerpt": "Hi, Huge fan of your work! Just to clarify, this project isn't an alternative to obsidian-cli. My goal is simply to enable semantic search across my vault by ingesting note chunks into ChromaDB while preserving all Wikilinks and tags. This allows AI agents to retrieve specific, semantically relevant text chunks based on user queries, rather than fetching entire notes. It also helps to find hidden connections between ", "context": "", "use_case": "semantic search over notes", "switched_from": "", "story": "Show HN: Local-first CLI to make Obsidian vaults searchable for AI agents"}, {"id": 780259058, "item_id": 48766191, "url": "https://news.ycombinator.com/item?id=48766191", "by": "emanuelecz", "date": "2026-07-02", "version": null, "stance": "neutral", "quote": "ChromaDB, pgvector", "excerpt": "Location: Basque Country, Spain (CET/CEST) Remote: Yes, remote only. EU contract, or US via EOR (Deel / Remote.com). Registered freelancer, can also invoice B2B. Willing to relocate: No Technologies: Python, LangGraph/LangChain, RAG (hybrid BM25 + dense retrieval, VoyageAI embeddings + reranking, ChromaDB, pgvector), LLM evals (golden sets, LLM-as-judge, judge meta-evaluation), Claude + OpenAI APIs, FastAPI, PostgreS", "context": "", "use_case": "vector search", "switched_from": "", "story": "Ask HN: Who wants to be hired? (July 2026)"}, {"id": 773918563, "item_id": 48369910, "url": "https://news.ycombinator.com/item?id=48369910", "by": "rasikjain", "date": "2026-06-02", "version": null, "stance": "neutral", "quote": "using LangChain, LangGraph, and vector databases like Pinecone and ChromaDB", "excerpt": "Location: Greater New York (US Citizen) Remote: Yes Willing to relocate: No About: I am a Generative AI Engineer who also happens to be pretty handy across the full stack. Most of my recent work has been building RAG pipelines, AI agents, and natural language search \u2014 using LangChain, LangGraph, and vector databases like Pinecone and ChromaDB \u2014 with OpenAI, Anthropic, and Gemini APIs under the hood. I care deeply abo", "context": "Generative AI Engineer", "use_case": "", "switched_from": "", "story": "Ask HN: Who wants to be hired? (June 2026)"}, {"id": 768242066, "item_id": 48015129, "url": "https://news.ycombinator.com/item?id=48015129", "by": "ravachol", "date": "2026-05-04", "version": null, "stance": "neutral", "quote": "Real-time ASCII visualizations (via Chroma)", "excerpt": "New in this version: New playback pipeline with improved performance and latency (built on miniaudio) Real-time ASCII visualizations (via Chroma) Free, open source, no tracking, no ads, completely offline. More info: https://codeberg.org/ravachol/kew Or: https://github.com/ravachol/kew", "context": "", "use_case": "", "switched_from": "", "story": "New kew v4.0 \"Love is gonna save us edition\" [video]"}, {"id": 766100320, "item_id": 47881270, "url": "https://news.ycombinator.com/item?id=47881270", "by": "Narew", "date": "2026-04-23", "version": null, "stance": "recommend", "quote": "I'm not colorblind but I use that quite often to check roughly if the color palette I choose is fine", "excerpt": "Ubisoft release some tools to simulate colorblindness. It's easy to extract de color transformation from their shader directly. I'm not colorblind but I use that quite often to check roughly if the color palette I choose is fine https://github.com/ubisoft/Chroma", "context": "not colorblind, designs palettes", "use_case": "colorblindness simulation", "switched_from": "", "story": "Your hex editor should color-code bytes"}, {"id": 760466448, "item_id": 47529153, "url": "https://news.ycombinator.com/item?id=47529153", "by": "nalinidash", "date": "2026-03-26", "version": null, "stance": "neutral", "quote": "ChromaDB is open source with Apache-2.0 license.", "excerpt": "ChromaDB is open source with Apache-2.0 license. https://github.com/chroma-core/chroma", "context": "", "use_case": "", "switched_from": "", "story": "From zero to a RAG system: successes and failures"}, {"id": 757739985, "item_id": 47358749, "url": "https://news.ycombinator.com/item?id=47358749", "by": "ptobey", "date": "2026-03-12", "version": null, "stance": "neutral", "quote": "local ChromaDB persistence", "excerpt": "I just open-sourced a project called Local Memory MCP: https://github.com/ptobey/local-memory-mcp It\u2019s a local-first personal memory layer for AI assistants: - text-first chunks + lightweight metadata - local ChromaDB persistence - MCP tools for store/search/update/delete/version chain - stdio and SSE transports - Docker-first setup I\u2019m looking for technical feedback from people building with MCP/RAG: 1. Does this ch", "context": "", "use_case": "", "switched_from": "", "story": "Feedback on a local-first MCP memory system for AI assistants?"}, {"id": 756919426, "item_id": 47307464, "url": "https://news.ycombinator.com/item?id=47307464", "by": "kristopolous", "date": "2026-03-09", "version": null, "stance": "neutral", "quote": "apparently I was still using chroma - I'll change that soon", "excerpt": "Cool to see this! I started a similar project in January but but nobody seemed interested in it at the time. Looks like I'll get back on that. https://github.com/day50-dev/infinite-mcp Essentially (1) start with the aggregator mcp repos: https://github.com/day50-dev/infinite-mcp/blob/main/gh-scrape.sh . pull all of them down. (2) get the meta information to understand how fresh, maintained, and popular the projects a", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: Mcp2cli \u2013 One CLI for every API, 96-99% fewer tokens than native MCP"}, {"id": 754637665, "item_id": 47164854, "url": "https://news.ycombinator.com/item?id=47164854", "by": "anulum", "date": "2026-02-26", "version": null, "stance": "neutral", "quote": "Your GroundTruthStore (ChromaDB) can be arbitrarily large, versioned, and updated", "excerpt": "@soletta Got it \u2014 thanks for the extra clarity, that\u2019s an important distinction. You\u2019re absolutely right: modern frontier models (Claude 3.5/Opus-class, GPT-4o, etc.) have become extremely good at maintaining internal consistency during autoregressive generation. They rarely contradict themselves within the same response anymore. Where Director-AI adds unique value is *external grounding + hard enforcement* against a", "context": "", "use_case": "Vector store for knowledge base", "switched_from": "", "story": "Show HN: Director-AI \u2013 token-level NLI+RAG"}, {"id": 754587186, "item_id": 47161699, "url": "https://news.ycombinator.com/item?id=47161699", "by": "anulum", "date": "2026-02-25", "version": null, "stance": "neutral", "quote": "The 0.4\u00d7 RAG weight pulls from your GroundTruthStore (ChromaDB).", "excerpt": "@soletta Great question \u2014 this is exactly why we built it this way. *Short answer*: frontier LLMs are excellent at static self-critique, but terrible for *real-time token-by-token streaming guardrails* because of latency, cost, and lack of persistent custom memory. *Why DeBERTa + RAG wins here*: - *Latency*: DeBERTa-v3-base + Rust kernel scores every ~4 tokens in ~220 ms (AggreFact eval). A frontier LLM call (GPT-4o/", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: Director-AI \u2013 token-level NLI+RAG"}, {"id": 746164368, "item_id": 46635273, "url": "https://news.ycombinator.com/item?id=46635273", "by": "philip1209", "date": "2026-01-15", "version": null, "stance": "neutral", "quote": "I've been using Chroma", "excerpt": "I run a Mac Mini home datacenter [1]. I've been using Chroma, Qwen 0.6B embeddings, and gpt-oss-20b to build a search agent over my blog. [1]: https://www.contraption.co/a-mini-data-center/", "context": "runs a Mac Mini home datacenter", "use_case": "Vector store for RAG search", "switched_from": "", "story": "Ask HN: How are you doing RAG locally?"}, {"id": 746115938, "item_id": 46632246, "url": "https://news.ycombinator.com/item?id=46632246", "by": "gaganyatri", "date": "2026-01-15", "version": null, "stance": "neutral", "quote": "ChromaDB for Vector storage.", "excerpt": "Built discovery using - Qwen-3-VL-8B for Document Ocr + Prompts + Tool Call - ChromaDB for Vector storage. - BM25 + Embedding model for Hybrid RAG. - Backend- FastAPI + Python - Frontend- React + Typescript - vllm + docker for model deployment on L40 GPU Demo: https://app.dwani.ai GitHub: https://github.com/dwani-ai/discovery Now working on added Agentic features, by continuous analysis of Document with Generated pro", "context": "", "use_case": "", "switched_from": "", "story": "Ask HN: How are you doing RAG locally?"}, {"id": 746102113, "item_id": 46631382, "url": "https://news.ycombinator.com/item?id=46631382", "by": "pj4533", "date": "2026-01-15", "version": null, "stance": "neutral", "quote": "storage in ChromaDB", "excerpt": "Well this isn\u2019t code, but I\u2019ve been working on a memory system for Claude Code. This portion provides semantic search over the session files in .claude/projects. It uses OpenAI for embeddings so not completely local (would be easy to modify) and storage in ChromaDB. https://github.com/pj4533/seance", "context": "", "use_case": "vector storage", "switched_from": "", "story": "Ask HN: How are you doing RAG locally?"}, {"id": 743980678, "item_id": 46498792, "url": "https://news.ycombinator.com/item?id=46498792", "by": "saeedq", "date": "2026-01-05", "version": null, "stance": "compare", "quote": "Only brinicle, and chroma survived", "excerpt": "My name is Saeed, and in general, I'm a deep learning specialist since 2019, but within the past 2 years, I've been writing optimized algorithms with cuda/C++. Recently, I decided to build a vector engine that consumes less resources than vector databases while not sacrificing accuracy, and throughput. So, we created brinicle. It is a production-oriented ANN index engine (not a full vector database) designed to stay ", "context": "", "use_case": "", "switched_from": "", "story": "I designed a vector engine to be RAM/CPU efficient"}, {"id": 743815809, "item_id": 46488488, "url": "https://news.ycombinator.com/item?id=46488488", "by": "brabel", "date": "2026-01-04", "version": null, "stance": "recommend", "quote": "Used this for code syntax higlighting: https://github.com/alecthomas/chroma", "excerpt": "I did the same thing, but implemented my site generator in Go. My site has grown by a lot over the years, but I can still build it from scratch (from MD files, HTML snippets and static files) in less than one second! Also have a RSS feed generator and it can highlight code in most programming languages, which is important to me as I write posts on many languages. I did try Hugo before I went on to implement my own, a", "context": "", "use_case": "code syntax highlighting", "switched_from": "", "story": "Jeffgeerling.com has been Migrated to Hugo"}, {"id": 741604514, "item_id": 46350282, "url": "https://news.ycombinator.com/item?id=46350282", "by": "kristopolous", "date": "2025-12-21", "version": null, "stance": "neutral", "quote": "database of my manpages with chroma", "excerpt": "I use local models + openrouter free ones. My monthly spend on ai models is < $1 I'm not cheap, just ahead of the curve. With the collapse in inference cost, everything will be this eventually I'll basically do $ man tool | <how do I do this with the tool> or even $ cat source | <find the flags and give me some documentation on how to use this> Things I used to do intensively I now do lazily. I've even made a IEITYua", "context": "", "use_case": "vector database for manpages", "switched_from": "", "story": "A guide to local coding models"}, {"id": 740590035, "item_id": 46286877, "url": "https://news.ycombinator.com/item?id=46286877", "by": "gardnr", "date": "2025-12-16", "version": null, "stance": "recommend", "quote": "I use Chroma for a lot of projects.", "excerpt": "pgVector is great and so is FAISS, but those are just a subset of what you get from Milvus. If all you need to do is RAG over 50Mb of documents then pick the right tool for the job. I use Chroma for a lot of projects. Then, what if you want hybrid search, or different IVF variants, or disk-based search, or horizontal scaling, or something that leverages SIMD, or sparse vectors? Milvus is great.", "context": "", "use_case": "", "switched_from": "", "story": "7 Years, 2 Rebuilds, 40K+ Stars: Milvus Recap and Roadmap"}, {"id": 740185937, "item_id": 46261621, "url": "https://news.ycombinator.com/item?id=46261621", "by": "NotLing", "date": "2025-12-14", "version": null, "stance": "neutral", "quote": "retrieves recent high-importance vectors from ChromaDB", "excerpt": "Hi HN, I'm a 17yo student from Taiwan. I was having a hard time with my multi-agent project, so I was like, what about we make LLM do whatever it wants instead? Maybe even self-evolve? So I built Atlas \u2013 an attempt to engineer a \"digital organism\" using Control Theory and Cognitive Science principles. The Core Architecture (How it works): 1. Adaptive Homeostasis (The Endocrine System) Instead of infinite loops, Atlas", "context": "", "use_case": "", "switched_from": "", "story": "Atlas \u2013 A Novel Artificial Intelligence Architecture"}, {"id": 740107009, "item_id": 46256688, "url": "https://news.ycombinator.com/item?id=46256688", "by": "helain", "date": "2025-12-13", "version": null, "stance": "neutral", "quote": "Ollama, Chroma/Qdrant", "excerpt": "If you don\u2019t want to reinvent all of this yourself, this is exactly the problem we\u2019re solving at Ailog. Most local LLM setups break down because people try to use the model as both the reasoning engine and the memory store. That doesn\u2019t scale. What works in production is a layered approach: external long-term memory (vector DB + metadata), short-term working state, aggressive summarization, and strict retrieval and e", "context": "", "use_case": "", "switched_from": "", "story": "Ask HN: How do you give a local AI model long-term memory?"}, {"id": 739143041, "item_id": 46196440, "url": "https://news.ycombinator.com/item?id=46196440", "by": "jellyotsiro", "date": "2025-12-08", "version": null, "stance": "recommend", "quote": "i use multiple DBs but the primary ones are turbopuffer and chroma for package search. they are really great", "excerpt": "hey! i use multiple DBs but the primary ones are turbopuffer and chroma for package search. they are really great re local, I do local for certain companies!", "context": "", "use_case": "Package search vector storage", "switched_from": "", "story": "Launch HN: Nia (YC S25) \u2013 Give better context to coding agents"}, {"id": 737394674, "item_id": 46087167, "url": "https://news.ycombinator.com/item?id=46087167", "by": "2dogsanerd", "date": "2025-11-29", "version": null, "stance": "neutral", "quote": "ChromaDB for vector storage", "excerpt": "Hi HN, I built this \"Community Edition\" kit because I wanted a clean, self-hosted way to chat with my documents AND my codebase without sending data to the cloud. It's a Docker Compose setup that wires together: Docling for parsing (handles PDFs/Tables really well) ChromaDB for vector storage FastAPI backend + Simple UI Ollama for local embeddings/LLM The main feature: It separates \"Code\" and \"Docs\" into different co", "context": "", "use_case": "vector storage", "switched_from": "", "story": "Show HN: Self-hosted RAG for docs and code (FastAPI, Docling, ChromaDB)"}, {"id": 736733922, "item_id": 46045870, "url": "https://news.ycombinator.com/item?id=46045870", "by": "acro-v", "date": "2025-11-25", "version": null, "stance": "neutral", "quote": "it implements embedding store locally - with ChromaDB and ONNXMiniLM_L6_V2 model", "excerpt": "Someone pointed me to this post from Cline engineer - below is my response to that Post: https://cline.bot/blog/why-cline-doesnt-index-your-codebase-and-why-thats-a-good-thing That post however does not apply to offline processing use case. Here are his 3 main problem points they re trying to solve: Code Doesn't Think in Chunks But then he is describing follow semantic links through imports, etc. -> that technique is", "context": "", "use_case": "", "switched_from": "", "story": "Can application layer improve local model output quality?"}, {"id": 736653152, "item_id": 46040822, "url": "https://news.ycombinator.com/item?id=46040822", "by": "ineedasername", "date": "2025-11-24", "version": null, "stance": "neutral", "quote": "summarized in a chromadb", "excerpt": "I needed some data from a content API, had a few options: 1) Human agent, manual retrieval (included for completion 2) one-off script to get exactly the content I want 3) Traditional workflow, write & maintain 4) one off prompt to the agent to write the script in #1, sort and arrange content for grouping base on descriptions it receives (this is what I used, 3 hours later I had a years worth of journal abstracts of v", "context": "", "use_case": "indexing and semantic search", "switched_from": "", "story": "An Economy of AI Agents"}, {"id": 736206961, "item_id": 46012935, "url": "https://news.ycombinator.com/item?id=46012935", "by": "ousamzing", "date": "2025-11-21", "version": null, "stance": "neutral", "quote": "RAG over the full codebase (ChromaDB + MiniLM embeddings)", "excerpt": "Creator here, thanks for checking this out. Code Intel is an experimental platform that combines: -AST parsing (to extract structure, anti-patterns, duplication), - Dependency graph analysis (circular imports, module cohesion), - RAG over the full codebase (ChromaDB + MiniLM embeddings), - Multi-agent LLM reasoning (Security/Performance/Architecture agents via GPT-4). All orchestrated via FastAPI + WebSockets, with a", "context": "", "use_case": "Vector store for RAG", "switched_from": "", "story": "Code Intel: Multi-agent LLM and AST analysis for Python codebases (Python only)"}]}