{"slug": "pgvector", "name": "pgvector", "category": "databases", "category_name": "Databases", "description": "", "recommenders": 13, "commenters": 47, "warners": 4, "score": 14.184, "rising": -2.93, "first": "2025-11-01", "latest": "2026-10-14", "why": "you can use quantized vector indexes[0] (handled by pgvector)", "why_is_quote": true, "has_verdict": false, "sentiment": {"recommend": 13, "neutral": 30, "compare": 3, "warn": 5}, "spark": [4, 1, 1, 3, 2, 0, 0, 1, 0, 1, 0, 0], "listed": true, "loved": [], "caveat": "", "short_quote": "you can use quantized vector indexes[0] (handled by pgvector)", "kind": "", "icon": false, "domain": null, "github_repo": "github.com/pgvector/pgvector", "official_url": "https://github.com/pgvector/pgvector", "pricing": "", "aliases": ["pgvector"], "verdict": null, "praise": [{"label": "", "summary": "", "quotes": [{"id": 789851873, "item_id": 49365742, "url": "https://news.ycombinator.com/item?id=49365742", "by": "KronisLV", "date": "2026-08-19", "stance": "recommend", "quote": "excellent plugins like pgvector and PostGIS", "excerpt": "> PostgreSQL allowed us to use a fulltext search plugin to do everything in one system. No need to sync any data. No need to maintain and run two systems. It just worked and made us smile (after some tweaks of course). Simplicity. I found MariaDB to be wonderfully simple to use for somewhat casual use cases: https://mariadb.com/docs/server/ha-and-performance/optimization-and-tuning/optimization-and-indexes/full-text-", "context": "", "use_case": "", "switched_from": "", "story": "PostgreSQL for Everything"}, {"id": 774844672, "item_id": 48427792, "url": "https://news.ycombinator.com/item?id=48427792", "by": "0xCMP", "date": "2026-06-06", "stance": "recommend", "quote": "you can use quantized vector indexes[0] (handled by pgvector)", "excerpt": "Some thing I've learned, but rarely seen explained anywhere: Storing the vectors is most likely not an issue, mostly likely you're having a problem with the indexes on top of them in which case you can use quantized vector indexes[0] (handled by pgvector) which will get past the limits imposed by PostgreSQL. I had to switch off pgvecto.rs at some point and figured that out. I don't have specific experience with the A", "context": "", "use_case": "vector search in PostgreSQL", "switched_from": "", "story": "pg_durable: Microsoft open sources in-database durable execution"}, {"id": 756757648, "item_id": 47297353, "url": "https://news.ycombinator.com/item?id=47297353", "by": "kencho", "date": "2026-03-08", "stance": "recommend", "quote": "pgvector is great if you're already on postgres and only need text search", "excerpt": "pgvector is great if you're already on postgres and only need text search. funny enough, Neon Postgres actually featured us in a case study about this exact topic. we replaced pinecone and rds with neon under the hood: https://neon.com/blog/vecstore-replacing-pinecone-and-rds-with-neon the gap shows up when you need image search, face search, or content moderation on top of text search. that's where a dedicated api m", "context": "", "use_case": "", "switched_from": "", "story": "You Don't Need a Vector Database"}, {"id": 755543552, "item_id": 47221472, "url": "https://news.ycombinator.com/item?id=47221472", "by": "dmix", "date": "2026-03-02", "stance": "recommend", "quote": "I just did that myself recently with pgvector for a documentation site.", "excerpt": "This is a good time to be offering hybrid search extensions. I just did that myself recently with pgvector for a documentation site. Does ParadeDB work with Render? They seem to have a whitelist of extensions https://render.com/docs/postgresql-extensions", "context": "", "use_case": "Vector search for docs site", "switched_from": "", "story": "Show HN: Omni \u2013 Open-source workplace search and chat, built on Postgres"}, {"id": 752883682, "item_id": 47055230, "url": "https://news.ycombinator.com/item?id=47055230", "by": "Arifcodes", "date": "2026-02-17", "stance": "recommend", "quote": "pgvector beats a dedicated vector DB for most use cases", "excerpt": "Go's tooling story in general is underrated. go fix, go vet, gopls - the stdlib toolchain does a lot that other ecosystems need third-party plugins for. On the subject of Go in production: I wrote up some lessons from running Go LLM pipelines at scale (fintech + gov workloads) if anyone is working on similar things. Covers caching strategies, error handling patterns specific to LLM failures, and why pgvector beats a ", "context": "", "use_case": "vector search", "switched_from": "", "story": "Using go fix to modernize Go code"}]}], "caveats": [{"label": "", "summary": "", "quotes": [{"id": 761437716, "item_id": 47589857, "url": "https://news.ycombinator.com/item?id=47589857", "by": "tjgreen", "date": "2026-03-31", "stance": "warn", "quote": "scaling semantic search beyond pgvector's main memory limitations", "excerpt": "Last summer we faced a conundrum at my company, Tiger Data, a Postgres cloud vendor whose main business is in timeseries data. We were trying to grow our business towards emerging AI-centric workloads and wanted to provide a state-of-the-art hybrid search stack in Postgres. We'd already built pgvectorscale in house with the goal of scaling semantic search beyond pgvector's main memory limitations. We just needed a sc", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: How This Graybeard Built the Fastest and Freest Postgres BM25 Search"}, {"id": 745927281, "item_id": 46620455, "url": "https://news.ycombinator.com/item?id=46620455", "by": "mips_avatar", "date": "2026-01-14", "stance": "warn", "quote": "pgvector had some perf problems", "excerpt": "Mostly usearch since pgvector had some perf problems, though I use postgres during my ingestion stage since postgis is so good. I segment my hnsw indexes by multipe h3 levels.", "context": "", "use_case": "", "switched_from": "", "story": "I built a geocoder for AI agents because I couldn't afford Google Maps"}, {"id": 732915024, "item_id": 45807189, "url": "https://news.ycombinator.com/item?id=45807189", "by": "ComputerGuru", "date": "2025-11-03", "stance": "warn", "quote": "pushing pgvector with very suspicious benchmarks that turned me off it altogether", "excerpt": "I had only heard positive things about pgvector but when you Google comparisons with leading vector dbs you keep getting seo slop from Tiger Data pushing pgvector with very suspicious benchmarks that turned me off it altogether instead https://www.tigerdata.com/blog/pgvector-vs-qdrant", "context": "", "use_case": "", "switched_from": "", "story": "The Case Against PGVector"}, {"id": 732790418, "item_id": 45799401, "url": "https://news.ycombinator.com/item?id=45799401", "by": "VoVAllen", "date": "2025-11-03", "stance": "warn", "quote": "solved most of the pgvector issues mentioned in this blog", "excerpt": "We at https://github.com/tensorchord/VectorChord solved most of the pgvector issues mentioned in this blog: - We're IVF + quantization, can support 15x more updates per second comparing to pgvector's HNSW. Insert or delete an element in a posting list is a super light operation comparing to modify a graph (HNSW) - Our main branch can now index 100M 768-dim vector in 20min with 16vcpu and 32G memory. This enables user", "context": "", "use_case": "", "switched_from": "", "story": "The Case Against PGVector"}]}], "who": [{"by": "lardissone", "context": "Senior full-stack developer, 25+ years experience", "url": "https://news.ycombinator.com/item?id=49924751", "date": "2026-10-01"}, {"by": "emanuelecz", "context": "self-taught AI engineer", "url": "https://news.ycombinator.com/item?id=49224408", "date": "2026-08-08"}, {"by": "Sriram1729", "context": "full-stack + AI engineer", "url": "https://news.ycombinator.com/item?id=48915249", "date": "2026-07-14"}, {"by": "yorktanaka2024", "context": "Senior software engineer, 10+ yrs", "url": "https://news.ycombinator.com/item?id=48768021", "date": "2026-07-02"}, {"by": "denizcakiroglu", "context": "full-stack engineer specializing in production LLM systems", "url": "https://news.ycombinator.com/item?id=48502797", "date": "2026-06-12"}, {"by": "m1r0k3", "context": "Runs recommendation system in production", "url": "https://news.ycombinator.com/item?id=45836468", "date": "2025-11-06"}, {"by": "jascha_eng", "context": "works at Discourse", "url": "https://news.ycombinator.com/item?id=45804121", "date": "2025-11-03"}], "tags": [], "use_cases": ["vector search", "vector similarity search", "vector search for rag", "vector search in postgres", "vector search in postgresql", "vector rag search"], "compared": [{"slug": "postgresql", "name": "PostgreSQL", "co": 26, "switched_from": 0, "switched_to": 0}, {"slug": "redis", "name": "Redis", "co": 7, "switched_from": 0, "switched_to": 0}, {"slug": "fastapi", "name": "FastAPI", "co": 6, "switched_from": 0, "switched_to": 0}, {"slug": "claude", "name": "Claude", "co": 6, "switched_from": 0, "switched_to": 0}, {"slug": "react", "name": "React", "co": 6, "switched_from": 0, "switched_to": 0}, {"slug": "langgraph", "name": "LangGraph", "co": 5, "switched_from": 0, "switched_to": 0}, {"slug": "docker", "name": "Docker", "co": 5, "switched_from": 0, "switched_to": 0}], "makers": [], "threads": [{"url": "https://news.ycombinator.com/item?id=45798479", "title": "The Case Against PGVector", "mentions": 11}, {"url": "https://news.ycombinator.com/item?id=48747975", "title": "Ask HN: Who wants to be hired? (July 2026)", "mentions": 3}, {"url": "https://news.ycombinator.com/item?id=49922568", "title": "Ask HN: Who wants to be hired? (October 2026)", "mentions": 2}, {"url": "https://news.ycombinator.com/item?id=49361279", "title": "PostgreSQL for Everything", "mentions": 2}, {"url": "https://news.ycombinator.com/item?id=49156682", "title": "Ask HN: Who wants to be hired? (August 2026)", "mentions": 2}, {"url": "https://news.ycombinator.com/item?id=48357724", "title": "Ask HN: Who wants to be hired? (June 2026)", "mentions": 2}], "sentiment_by_month": {"2025-11": {"neutral": 3, "recommend": 4, "warn": 3}, "2025-12": {"compare": 2, "recommend": 1, "neutral": 2}, "2026-01": {"neutral": 4, "warn": 1, "recommend": 1}, "2026-02": {"recommend": 3, "neutral": 2, "compare": 1}, "2026-03": {"recommend": 2, "neutral": 4, "warn": 1}, "2026-04": {}, "2026-05": {}, "2026-06": {"neutral": 3, "recommend": 1}, "2026-07": {"neutral": 6}, "2026-08": {"neutral": 3, "recommend": 1}, "2026-09": {"neutral": 1}, "2026-10": {"neutral": 2}}, "quotes": [{"id": 801407430, "item_id": 50087964, "url": "https://news.ycombinator.com/item?id=50087964", "by": "akshit_29", "date": "2026-10-14", "stance": "neutral", "quote": "pgvector RAG", "excerpt": "Location: Gurugram, India (IST, UTC+5:30) Remote: Yes, remote only. 4-6h daily overlap with US Eastern, full overlap with EU. Contractor or EOR. Willing to relocate: No Technologies: Python, Django/DRF, FastAPI, Celery, Kafka, PostgreSQL, Redis, Next.js, LangGraph, OpenAI Agents SDK, MCP, pgvector, Pipecat, AWS, Azure, Kubernetes Resume/CV: https://akshit.wtf Email: ahujaakshit20@gmail.com Backend engineer, around 4 ", "context": "", "use_case": "vector RAG search", "switched_from": "", "story": "Ask HN: Who wants to be hired? (October 2026)"}, {"id": 798796016, "item_id": 49924751, "url": "https://news.ycombinator.com/item?id=49924751", "by": "lardissone", "date": "2026-10-01", "stance": "neutral", "quote": "using Python, pgvector, and OpenAI, Claude, and Gemini APIs", "excerpt": "Location: Argentina (US Eastern timezone overlap) Remote: Yes Willing to relocate: No Technologies: Python, Django, Node.js, NestJS, TypeScript, React, Next.js, React Native, PostgreSQL, Redis, AWS, Vercel, Docker, AI/LLMs, RAG, pgvector, AI agents R\u00e9sum\u00e9/CV: https://linkedin.com/in/lardissone | https://leandroardissone.com/resume Email: leandro@ardissone.com I'm a senior full-stack developer with 25+ years of experi", "context": "Senior full-stack developer, 25+ years experience", "use_case": "Vector search for RAG", "switched_from": "", "story": "Ask HN: Who wants to be hired? (October 2026)"}, {"id": 792390488, "item_id": 49524405, "url": "https://news.ycombinator.com/item?id=49524405", "by": "sjtroxel", "date": "2026-09-01", "stance": "neutral", "quote": "pgvector, PostgreSQL", "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": 789941505, "item_id": 49371344, "url": "https://news.ycombinator.com/item?id=49371344", "by": "jankovicsandras", "date": "2026-08-20", "stance": "neutral", "quote": "plpgsql_bm25 + pgvector", "excerpt": "You can do BM25 and hybrid search in Postgres. Shameless plug: https://github.com/jankovicsandras/plpgsql_bm25 BM25 search implemented in PL/pgSQL ( Unlicense / Public domain ) The repo includes also plpgsql_bm25rrf.sql : PL/pgSQL function for hybrid search ( plpgsql_bm25 + pgvector ) with Reciprocal Rank Fusion; and Jupyter notebook examples.", "context": "", "use_case": "", "switched_from": "", "story": "PostgreSQL for Everything"}, {"id": 789851873, "item_id": 49365742, "url": "https://news.ycombinator.com/item?id=49365742", "by": "KronisLV", "date": "2026-08-19", "stance": "recommend", "quote": "excellent plugins like pgvector and PostGIS", "excerpt": "> PostgreSQL allowed us to use a fulltext search plugin to do everything in one system. No need to sync any data. No need to maintain and run two systems. It just worked and made us smile (after some tweaks of course). Simplicity. I found MariaDB to be wonderfully simple to use for somewhat casual use cases: https://mariadb.com/docs/server/ha-and-performance/optimization-and-tuning/optimization-and-indexes/full-text-", "context": "", "use_case": "", "switched_from": "", "story": "PostgreSQL for Everything"}, {"id": 787590531, "item_id": 49224408, "url": "https://news.ycombinator.com/item?id=49224408", "by": "emanuelecz", "date": "2026-08-08", "stance": "neutral", "quote": "Postgres/pgvector", "excerpt": "Location: Gipuzkoa, Spain (CET) Remote: Yes, remote only Willing to relocate: No Technologies: Python, FastAPI, LangGraph, MCP, RAG, Postgres/pgvector, Redis, Celery, Docker, LLM evals, TypeScript, Next.js R\u00e9sum\u00e9/CV: https://www.emanuelecz.dev/emanuele-czofei-cv.pdf Email: emanuelecz06@gmail.com Self-taught AI engineer, two years building, four months at a startup. Strongest at the unglamorous half: retrieval that re", "context": "self-taught AI engineer", "use_case": "vector storage", "switched_from": "", "story": "Ask HN: Who wants to be hired? (August 2026)"}, {"id": 786512036, "item_id": 49157002, "url": "https://news.ycombinator.com/item?id=49157002", "by": "tomxmp", "date": "2026-08-03", "stance": "neutral", "quote": "pgvector", "excerpt": "Location: Lisbon, Portugal (GMT+0/+1) Remote: Yes Willing to relocate: No Technologies: Typescript, React, TanStack Start, Next.js, PostgreSQL, pgvector, Redis, Docker, Drizzle ORM, Zod, Zustand, TanStack Query, Tailwind CSS, Plasmo, Vite, Astro, Expo, AI SDK, Python, TensorFlow, scikit-learn, Pandas, Deck.gl, AWS, GCP R\u00e9sum\u00e9/CV: https://tomasmenezes.com/cv Email: tomas.alexmp+hn@gmail.com More info: https://tomasmen", "context": "", "use_case": "", "switched_from": "", "story": "Ask HN: Who wants to be hired? (August 2026)"}, {"id": 782643988, "item_id": 48915249, "url": "https://news.ycombinator.com/item?id=48915249", "by": "Sriram1729", "date": "2026-07-14", "stance": "neutral", "quote": "Postgres + pgvector", "excerpt": "Tampa, FL | Remote (US Eastern) | Contract, ~20 hrs/week (evenings + weekends, around a full-time job)I'm a full-stack + AI engineer, 5 years in. Looking for one side engagement I can actually commit to \u2014 nights and weekends, not a flaky \"I'll get to it\" thing. Short version of what I'm good for: give me a problem and I'll hand back something deployed. I do the whole thing \u2014 backend, frontend, the data model, the inf", "context": "full-stack + AI engineer", "use_case": "vector search", "switched_from": "", "story": "Ask HN: Who wants to be hired? (July 2026)"}, {"id": 781456128, "item_id": 48841008, "url": "https://news.ycombinator.com/item?id=48841008", "by": "karmakaze", "date": "2026-07-08", "stance": "neutral", "quote": "turn the documents into vectors in say PostgreSQL (with pgvector)", "excerpt": "I learned a solution is to turn the documents into vectors in say PostgreSQL (with pgvector) and do a cosine similarity search with a search vector. Doing a search for embed models on HuggingFace shows nomic-ai/nomic-embed-text-v1.5 and Qwen/Qwen3-Embedding-0.6B. I might have used a larger one like Qwen/Qwen3-Embedding-4B. There's some info for AnythingLLM[0] which supports RAG. AnythingLLM has LanceDB out of the box", "context": "", "use_case": "Vector search", "switched_from": "", "story": "Turning a pile of documents into a searchable useable knowledge base"}, {"id": 781186402, "item_id": 48824150, "url": "https://news.ycombinator.com/item?id=48824150", "by": "daleverett", "date": "2026-07-07", "stance": "neutral", "quote": "PgVector already exists", "excerpt": "Hi HN, I\u2019m building Polygres. TLDR: Polygres turns Postgres into a full graph-rag database for agents by combining relational data, graph traversal, vector search, hybrid search, and ranking over the same source-of-truth data. Insert via sql, query as a graph, and vice versa. PgVector already exists, but we added the missing component which was PgGraph (OSS). You can check that out here: https://github.com/Evokoa/pgG", "context": "", "use_case": "vector search in Postgres", "switched_from": "", "story": "Show HN: Instant GraphRAG over any Postgres database"}, {"id": 780548513, "item_id": 48784282, "url": "https://news.ycombinator.com/item?id=48784282", "by": "rkochanowski", "date": "2026-07-04", "stance": "neutral", "quote": "I used this in a different project (halfvec in pgvector) without affecting results.", "excerpt": "Thanks, I will look at this in more detail. To give more context, the current version is already an optimized one I considered good enough and didn't spend more time on it. In the first attempt, I used a vector database with indexes, trying to query for similar vectors. This was uselessly slow even in medium-sized repos. The brute-force NumPy solution is a significant improvement, making it faster than other calculat", "context": "", "use_case": "vector storage with 16-bit floats", "switched_from": "", "story": "Show HN: CLI tool for detecting non-exact code duplication with embedding models"}, {"id": 780288344, "item_id": 48768021, "url": "https://news.ycombinator.com/item?id=48768021", "by": "yorktanaka2024", "date": "2026-07-02", "stance": "neutral", "quote": "pgVector", "excerpt": "Location: Windsor, ON N8Y 3J8, Canada Remote: Yes Willing to relocate: Yes Technologies: AI(LLMs, Multi-Agent workflow, Automation, Langchain, LangGraph, LangSmith, RAG, MCP) + Full Stack(Typescript, Python, .NET) + DevOps/Cloud Infra(AWS, GCP, Terraform, CI/CD) + RabbitMQ/Kafka + Redis + PostgreSQL/pgVector + Microservices R\u00e9sum\u00e9/CV: https://yorktanaka.online/cv Email: york.tiger.2024@gmail.com Senior Software Engin", "context": "Senior software engineer, 10+ yrs", "use_case": "Vector search", "switched_from": "", "story": "Ask HN: Who wants to be hired? (July 2026)"}, {"id": 780259059, "item_id": 48766191, "url": "https://news.ycombinator.com/item?id=48766191", "by": "emanuelecz", "date": "2026-07-02", "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": 776044764, "item_id": 48502797, "url": "https://news.ycombinator.com/item?id=48502797", "by": "denizcakiroglu", "date": "2026-06-12", "stance": "neutral", "quote": "", "excerpt": "Location: Istanbul, T\u00fcrkiye (relocating to Guildford, UK in Sept 2026) Remote: Yes \u2014 full EU overlap; US-East evenings (UTC+3, moving to UTC+0) Willing to relocate: No Technologies: TypeScript, Node.js, NestJS, Fastify, Next.js, React, React Native/Expo, PostgreSQL/Prisma, Redis/BullMQ, Claude API, RAG, tool-calling/MCP, pgvector, Playwright R\u00e9sum\u00e9/CV: https://github.com/denizcakiroglu322 Email: denizcakiroglu322@gma", "context": "full-stack engineer specializing in production LLM systems", "use_case": "vector search for RAG", "switched_from": "", "story": "Ask HN: Who wants to be hired? (June 2026)"}, {"id": 774844672, "item_id": 48427792, "url": "https://news.ycombinator.com/item?id=48427792", "by": "0xCMP", "date": "2026-06-06", "stance": "recommend", "quote": "you can use quantized vector indexes[0] (handled by pgvector)", "excerpt": "Some thing I've learned, but rarely seen explained anywhere: Storing the vectors is most likely not an issue, mostly likely you're having a problem with the indexes on top of them in which case you can use quantized vector indexes[0] (handled by pgvector) which will get past the limits imposed by PostgreSQL. I had to switch off pgvecto.rs at some point and figured that out. I don't have specific experience with the A", "context": "", "use_case": "vector search in PostgreSQL", "switched_from": "", "story": "pg_durable: Microsoft open sources in-database durable execution"}, {"id": 774363445, "item_id": 48397715, "url": "https://news.ycombinator.com/item?id=48397715", "by": "tjsawyer", "date": "2026-06-04", "stance": "neutral", "quote": "pgvector", "excerpt": "Location: Decatur, GA, USA Remote: Yes (remote only) Willing to relocate: No Technologies: Node.js, Express, PostgreSQL/Supabase, pgvector, React, React Router v7, Tailwind, Vercel, Render. Claude (Anthropic), Gemini 3 (Vertex), Kimi K2.6 (Groq), Llama (Groq), Voyage 3 embeddings, ElevenLabs TTS. R\u00e9sum\u00e9/CV: https://www.travissawyer.io/ Portfolio: https://www.regild.ai | Origin story: https://regild.ai/blog/origin-sto", "context": "", "use_case": "", "switched_from": "", "story": "Ask HN: Who wants to be hired? (June 2026)"}, {"id": 773741843, "item_id": 48358865, "url": "https://news.ycombinator.com/item?id=48358865", "by": "CharlieDigital", "date": "2026-06-01", "stance": "neutral", "quote": "", "excerpt": "> ...add something like apache AGE, but arguably that is also a small ecosystem (at least IMHO as I never heard it until I actually started looking for Neo4J alternatives) Outside of the most trivial use cases, I've found that AGE will not get anywhere near Neo4j in terms of performance and there's a lot of edge cases that just flat out won't work. The interesting types of queries you'd want to do in the graph end up", "context": "", "use_case": "", "switched_from": "", "story": "Benchmarking SurrealDB 3.x vs. Postgres, Mongo, Neo4j and Redis (With Fsync)"}, {"id": 761437716, "item_id": 47589857, "url": "https://news.ycombinator.com/item?id=47589857", "by": "tjgreen", "date": "2026-03-31", "stance": "warn", "quote": "scaling semantic search beyond pgvector's main memory limitations", "excerpt": "Last summer we faced a conundrum at my company, Tiger Data, a Postgres cloud vendor whose main business is in timeseries data. We were trying to grow our business towards emerging AI-centric workloads and wanted to provide a state-of-the-art hybrid search stack in Postgres. We'd already built pgvectorscale in house with the goal of scaling semantic search beyond pgvector's main memory limitations. We just needed a sc", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: How This Graybeard Built the Fastest and Freest Postgres BM25 Search"}, {"id": 757442785, "item_id": 47340174, "url": "https://news.ycombinator.com/item?id=47340174", "by": "DrMeric", "date": "2026-03-11", "stance": "neutral", "quote": "vector similarity (pgvector) run in parallel", "excerpt": "Orthopedics is inherently translational. A single clinical question can span surgical technique, implant biomechanics, material science, and computational modeling. Existing search engines rely on keyword matching or citation-based ranking. Neither approach handles cross-disciplinary queries well. A search for \"PEEK versus titanium spinal fusion cage mechanical properties\" returns thousands of irrelevant clinical res", "context": "", "use_case": "Vector similarity search", "switched_from": "", "story": "OrthoScience \u2013 Hybrid search engine for 500K+ orthopedic translational research"}, {"id": 757168021, "item_id": 47323001, "url": "https://news.ycombinator.com/item?id=47323001", "by": "pedro-freitas", "date": "2026-03-10", "stance": "neutral", "quote": "Technologies: Python, FastAPI, n8n, CrewAI, LangGraph, LangChain, RAG/pgvector, WhatsApp (Botmaker/Twilio), PostgreSQL, Docker, Coolify/DigitalOcean", "excerpt": "SEEKING WORK | Remote | AI Agents, Automation & Vertical AI Solutions Location: Rio de Janeiro, Brazil | Remote worldwide Technologies: Python, FastAPI, n8n, CrewAI, LangGraph, LangChain, RAG/pgvector, WhatsApp (Botmaker/Twilio), PostgreSQL, Docker, Coolify/DigitalOcean We're PD7 Lab \u2014 an AI lab and software factory building vertical AI solutions and custom software for companies in Brazil and globally. What we build", "context": "", "use_case": "", "switched_from": "", "story": "Ask HN: Freelancer? Seeking freelancer? (March 2026)"}, {"id": 757100436, "item_id": 47318777, "url": "https://news.ycombinator.com/item?id=47318777", "by": "chunktort", "date": "2026-03-09", "stance": "neutral", "quote": "pgvector + Supabase", "excerpt": "Hi \u2014 this stack is my daily driver. I built Jorge, a live FastAPI + Redis + Claude AI real estate qualification bot (1,753 passing tests, deployed on Render) that handles inbound SMS intake, multi-turn qualification, calendar booking, and human-handoff logic via GHL CRM. I also built DocExtract AI (FastAPI + ARQ workers + pgvector + Supabase, live at docextract-api.onrender.com) and an agentic workflow API with SSE s", "context": "", "use_case": "", "switched_from": "", "story": "Ask HN: Who is hiring? (March 2026)"}, {"id": 756881569, "item_id": 47305098, "url": "https://news.ycombinator.com/item?id=47305098", "by": "kaizenb", "date": "2026-03-08", "stance": "neutral", "quote": "pgvector cosine similarity on 768-dim embeddings", "excerpt": "Designing a conversational UX for Bookmarker. I was stuck on this conversation problem. First version had a dead-end search box: six starter prompts, one referencing a tool that didn't exist. No follow-ups. No guided flows. Users got an answer and had to invent the next question from scratch. Now the assistant explores your library with you. Tag discovery, color browsing, weekly digests, smart collections that auto-c", "context": "", "use_case": "Vector similarity search", "switched_from": "", "story": "Ask HN: What Are You Working On? (March 2026)"}, {"id": 756757648, "item_id": 47297353, "url": "https://news.ycombinator.com/item?id=47297353", "by": "kencho", "date": "2026-03-08", "stance": "recommend", "quote": "pgvector is great if you're already on postgres and only need text search", "excerpt": "pgvector is great if you're already on postgres and only need text search. funny enough, Neon Postgres actually featured us in a case study about this exact topic. we replaced pinecone and rds with neon under the hood: https://neon.com/blog/vecstore-replacing-pinecone-and-rds-with-neon the gap shows up when you need image search, face search, or content moderation on top of text search. that's where a dedicated api m", "context": "", "use_case": "", "switched_from": "", "story": "You Don't Need a Vector Database"}, {"id": 755543552, "item_id": 47221472, "url": "https://news.ycombinator.com/item?id=47221472", "by": "dmix", "date": "2026-03-02", "stance": "recommend", "quote": "I just did that myself recently with pgvector for a documentation site.", "excerpt": "This is a good time to be offering hybrid search extensions. I just did that myself recently with pgvector for a documentation site. Does ParadeDB work with Render? They seem to have a whitelist of extensions https://render.com/docs/postgresql-extensions", "context": "", "use_case": "Vector search for docs site", "switched_from": "", "story": "Show HN: Omni \u2013 Open-source workplace search and chat, built on Postgres"}, {"id": 754922338, "item_id": 47182646, "url": "https://news.ycombinator.com/item?id=47182646", "by": "alfredoizjr", "date": "2026-02-27", "stance": "neutral", "quote": "pgvector for semantic search", "excerpt": "Thanks! We use PostgreSQL (Supabase) with pgvector. Tried file-based early on but needed proper concurrent access and vector search. Indexing is hybrid \u2014 pgvector for semantic search, full-text search for exact matches. For lifecycle, we keep it simple: users organize knowledge into \"spaces\" per project. Git spaces auto-sync commits/PRs. No automatic eviction \u2014 snapshots let you save/restore state. Cross-conversation", "context": "", "use_case": "vector search", "switched_from": "", "story": "Show HN: ContextForge \u2013 Persistent memory MCP server for Claude"}, {"id": 752883682, "item_id": 47055230, "url": "https://news.ycombinator.com/item?id=47055230", "by": "Arifcodes", "date": "2026-02-17", "stance": "recommend", "quote": "pgvector beats a dedicated vector DB for most use cases", "excerpt": "Go's tooling story in general is underrated. go fix, go vet, gopls - the stdlib toolchain does a lot that other ecosystems need third-party plugins for. On the subject of Go in production: I wrote up some lessons from running Go LLM pipelines at scale (fintech + gov workloads) if anyone is working on similar things. Covers caching strategies, error handling patterns specific to LLM failures, and why pgvector beats a ", "context": "", "use_case": "vector search", "switched_from": "", "story": "Using go fix to modernize Go code"}, {"id": 752516017, "item_id": 47032251, "url": "https://news.ycombinator.com/item?id=47032251", "by": "miga", "date": "2026-02-16", "stance": "compare", "quote": "Why no benchmarks against pg_vector, DuckDB with extension?", "excerpt": "Why no benchmarks against pg_vector, DuckDB with extension? For benchmarks you may just prepare Phoronix Test Suite module (https://www.phoronix-test-suite.com/) to facilitate replication on a variety of machines.", "context": "", "use_case": "", "switched_from": "", "story": "zvec: embedded vector database"}, {"id": 751894225, "item_id": 46993389, "url": "https://news.ycombinator.com/item?id=46993389", "by": "readme", "date": "2026-02-12", "stance": "recommend", "quote": "i love postgres and pgvector... this is exactly to my tastes", "excerpt": "i love postgres and pgvector... this is exactly to my tastes", "context": "", "use_case": "Vector search extension", "switched_from": "", "story": null}, {"id": 751710690, "item_id": 46981918, "url": "https://news.ycombinator.com/item?id=46981918", "by": "pstryder", "date": "2026-02-11", "stance": "neutral", "quote": "PostgreSQL + pgvector for production", "excerpt": "# Show HN: Cathedral \u2013 Self-hosted AI chat with persistent memory, tools, and RAG *Title options:* 1. Show HN: Cathedral \u2013 Self-hosted AI chat with persistent memory, 31 tools, and document RAG 2. Show HN: Cathedral \u2013 A local AI operating system with memory that survives sessions 3. Show HN: Cathedral \u2013 Self-hosted, memory-augmented AI chat that works with any LLM --- *Post body:* I wanted an AI setup where nothing g", "context": "", "use_case": "", "switched_from": "", "story": "Self-hosted, memory-augmented AI chat that works with any LLM"}, {"id": 749574851, "item_id": 46848428, "url": "https://news.ycombinator.com/item?id=46848428", "by": "cckolon", "date": "2026-02-01", "stance": "recommend", "quote": "have seen very fast lookups with pgvector (HNSW index) and sqlite-vec on 20k vectors", "excerpt": "I use full length vectors (512 dimensions) and have seen very fast lookups with pgvector (HNSW index) and sqlite-vec on 20k vectors. I think any decent vector database should be able to handle 44k entries\u2026 which one are you using now?", "context": "", "use_case": "", "switched_from": "", "story": "Tuning Semantic Search on Jfmm.net"}, {"id": 748851617, "item_id": 46803226, "url": "https://news.ycombinator.com/item?id=46803226", "by": "gomoboo", "date": "2026-01-28", "stance": "neutral", "quote": "It\u2019s built with ASP.NET, Postgres/pgvector, and OpenAI embedding/LLMs.", "excerpt": "I love reading battlefield notes like this for RAG/search systems. Anyone shooting for useful output is going to hit the same pain points but each article like this has a different set of solutions. I\u2019m leaning on OpenAI for my embedding needs but will be trying llama-server in the future. I stuck with Postgres because it was easy to run it on my Dokku installation. Great to know sqlite is an option there too. My cor", "context": "", "use_case": "vector search", "switched_from": "", "story": "Tuning Semantic Search on Jfmm.net"}, {"id": 746143104, "item_id": 46633944, "url": "https://news.ycombinator.com/item?id=46633944", "by": "cipherself", "date": "2026-01-15", "stance": "recommend", "quote": "You can use pgvector for the vector lookup and paradedb for bm25.", "excerpt": "Here's a Dockerfile that will spin up postgres with pgvector and paradedb https://gist.github.com/cipherself/5260fea1e2631e9630081fb7db171598 You can use pgvector for the vector lookup and paradedb for bm25.", "context": "", "use_case": "Vector search", "switched_from": "", "story": "Ask HN: How are you doing RAG locally?"}, {"id": 746117681, "item_id": 46632355, "url": "https://news.ycombinator.com/item?id=46632355", "by": "juanre", "date": "2026-01-15", "stance": "neutral", "quote": "It uses PostgreSQL with pgvector, hybrid BM25, multi-query expansion, and reranking.", "excerpt": "I built https://github.com/juanre/llmemory and I use it both locally and as part of company apps. Quite happy with the performance. It uses PostgreSQL with pgvector, hybrid BM25, multi-query expansion, and reranking. (It's the first time I share it publicly, so I am sure there'll be quirks.)", "context": "", "use_case": "vector search", "switched_from": "", "story": "Ask HN: How are you doing RAG locally?"}, {"id": 745927281, "item_id": 46620455, "url": "https://news.ycombinator.com/item?id=46620455", "by": "mips_avatar", "date": "2026-01-14", "stance": "warn", "quote": "pgvector had some perf problems", "excerpt": "Mostly usearch since pgvector had some perf problems, though I use postgres during my ingestion stage since postgis is so good. I segment my hnsw indexes by multipe h3 levels.", "context": "", "use_case": "", "switched_from": "", "story": "I built a geocoder for AI agents because I couldn't afford Google Maps"}, {"id": 745572179, "item_id": 46598261, "url": "https://news.ycombinator.com/item?id=46598261", "by": "ArkhamMirror", "date": "2026-01-12", "stance": "neutral", "quote": "pgvector for embeddings", "excerpt": "I hate that we have to expose data to the cloud in order to analyze it, so I decided to build a platform that would help people get to the truth of things without having anyone peeking over their shoulder. ArkhamMirror: SHATTERED is a modular, air-gappable, local-first and AI-assisted platform. It can run entirely on your machine - no cloud, no telemetry, no ads, no subscriptions, no fees, no limits. It\u2019s designed fo", "context": "", "use_case": "", "switched_from": "", "story": "ArkhamMirror SHATTERED: Air-gapped investigative analysis, no Palantir required"}, {"id": 745125857, "item_id": 46570366, "url": "https://news.ycombinator.com/item?id=46570366", "by": "hising", "date": "2026-01-10", "stance": "neutral", "quote": "using PostgreSQL with pgvector", "excerpt": "Yeah, I had a similar idea, I used Open AI API to break down movies into the 3 act structure, narrative, pacing, character arcs etc and then trying to find movies that are similar using PostgreSQL with pgvector. The idea was to have another way to find movies I would like to watch next based on more than \"similar movies\" in IMDb. Threw some hours at it, but I guess it is a system that needs a lot of data, a lot of to", "context": "", "use_case": "vector search", "switched_from": "", "story": "Show HN: I used Claude Code to discover connections between 100 books"}, {"id": 743018992, "item_id": 46438687, "url": "https://news.ycombinator.com/item?id=46438687", "by": "jascha_eng", "date": "2025-12-30", "stance": "neutral", "quote": "combine the pgvector score and the bm25 score", "excerpt": "We have docs on how to do hybrid search here: https://www.tigerdata.com/docs/use-timescale/latest/extensions/pg-textsearch#build-hybrid-search-with-semantic-and-keyword-search Essentially you combine the pgvector score and the bm25 score to hopefully get better results.", "context": "", "use_case": "vector search", "switched_from": "", "story": "Postgres extension complements pgvector for performance and scale"}, {"id": 742970896, "item_id": 46435681, "url": "https://news.ycombinator.com/item?id=46435681", "by": "isoprophlex", "date": "2025-12-30", "stance": "compare", "quote": "comparing well-tuned pgvector to pinecone", "excerpt": "The linked blogpost is an interesting read, too, comparing well-tuned pgvector to pinecone: https://www.tigerdata.com/blog/pgvector-vs-pinecone", "context": "", "use_case": "", "switched_from": "", "story": "Postgres extension complements pgvector for performance and scale"}, {"id": 740623554, "item_id": 46288972, "url": "https://news.ycombinator.com/item?id=46288972", "by": "jankovicsandras", "date": "2025-12-16", "stance": "neutral", "quote": "plpgsql_bm25 + pgvector", "excerpt": "You can do hybrid search in Postgres. Shameless plug: https://github.com/jankovicsandras/plpgsql_bm25 BM25 search implemented in PL/pgSQL ( Unlicense / Public domain ) The repo includes plpgsql_bm25rrf.sql : PL/pgSQL function for hybrid search ( plpgsql_bm25 + pgvector ) with Reciprocal Rank Fusion; and Jupyter notebook examples.", "context": "", "use_case": "Vector search", "switched_from": "", "story": "7 Years, 2 Rebuilds, 40K+ Stars: Milvus Recap and Roadmap"}, {"id": 740590032, "item_id": 46286877, "url": "https://news.ycombinator.com/item?id=46286877", "by": "gardnr", "date": "2025-12-16", "stance": "recommend", "quote": "pgVector is great", "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"}]}