--- title: CommSync Runs Text, Vector, and Hybrid Search on... description: CommSync, a unified inbox platform for business communications, uses Neon's Lakebase Search extensions (lakebase_text and lakebase_vector) to power three... canonical: https://daily.dev/posts/commsync-runs-text-vector-and-hybrid-search-on-postgres-with-lakebase-search-sybz5is1d twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: CommSync Runs Text, Vector, and Hybrid Search on Postgres with Lakebase Search | daily.dev og:description: CommSync, a unified inbox platform for business communications, uses Neon's Lakebase Search extensions (lakebase_text and lakebase_vector) to power three... og:url: https://daily.dev/posts/commsync-runs-text-vector-and-hybrid-search-on-postgres-with-lakebase-search-sybz5is1d og:image: https://api.daily.dev/og/posts/SyBz5iS1D.png og:image:alt: CommSync Runs Text, Vector, and Hybrid Search on Postgres with Lakebase Search og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # CommSync Runs Text, Vector, and Hybrid Search on Postgres with Lakebase Search **[Neon](https://daily.dev/sources/neontech)** · 5 min read · 0 upvotes · 0 comments ## Summary CommSync, a unified inbox platform for business communications, uses Neon's Lakebase Search extensions (lakebase_text and lakebase_vector) to power three distinct search surfaces entirely within Postgres. Their main inbox uses BM25 text search via lakebase_text, achieving 18.6ms warm latency versus 19.5 seconds with the old GIN approach — a 1,000x improvement. An AI assistant uses vector search with 1024-dimension embeddings and cosine similarity. A new command palette combines both via Reciprocal Rank Fusion (RRF) with a two-tier approach: fast lexical-only pass (~110ms) followed by a debounced semantic pass (~420ms). Additionally, CommSync Agents use lakebase_vector for RAG over an org-level knowledge base, with tenancy and permissions handled as SQL WHERE clauses — no separate vector database required. The setup allows choosing between approximate nearest-neighbor (ANN) indexes for large corpora and exact nearest-neighbor for smaller ones, all within a single Postgres database. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is lakebase_text and how does it improve full-text search compared to native Postgres GIN indexes? lakebase_text is a Postgres extension bundled in Neon's Lakebase Search that adds a lakebase_bm25 index type for full-text search. It works with standard tsvector types and operators but adds true BM25 relevance scoring and top-K pushdown, features native GIN indexes lack. One customer reported query latency dropping from about 19.5 seconds cold-start with GIN/tsquery to about 18.6ms warm. _daily.dev surfaces real-world benchmarks like this for teams deciding between Postgres-native and extended full-text search._ ### Is lakebase_vector compatible with existing pgvector queries? Yes, lakebase_vector is a drop-in replacement for pgvector, using the same types, distance operators, and query syntax, so existing pgvector queries work unchanged. It adds the lakebase_ann index type for approximate nearest-neighbor search, while exact nearest-neighbor search with no index remains an option for small datasets needing perfect recall. _Track Postgres vector search extensions like this on daily.dev before migrating a pgvector-based RAG pipeline._ ### How can I combine keyword and semantic search into a single hybrid search ranking? One production approach uses Reciprocal Rank Fusion (RRF) with the formula score = sum of 1/(k + rank) across a lexical arm and a semantic arm, using k=60. Weighting the lexical arm at 1.0 and the semantic arm at 0.8 lets exact keyword matches win ties over conceptually similar but non-literal semantic hits, useful when keyword traffic dominates. _daily.dev helps developers comparing hybrid search ranking strategies keep up with real implementation details._ ## Similar posts on daily.dev - [Lakebase Search: State-of-the-art full text and vector search for Postgres](https://daily.dev/posts/lakebase-search-state-of-the-art-full-text-and-vector-search-for-postgres-gafr152pe) · databricks · 0 upvotes · 0 comments - [Lakebase Search: vector and BM25 on Neon](https://daily.dev/posts/lakebase-search-vector-and-bm25-on-neon-lvk6nbt3l) · Neon · 4 upvotes · 0 comments - [Announcing Lakebase Search: agent-native retrieval built into Lakebase Postgres](https://daily.dev/posts/announcing-lakebase-search-agent-native-retrieval-built-into-lakebase-postgres-uqtg9rcz2) · databricks · 1 upvotes · 0 comments - [Timescale Vector x LangChain: Making PostgreSQL A Better Vector Database for AI Applications](https://daily.dev/posts/timescale-vector-x-langchain-making-postgresql-a-better-vector-database-for-ai-applications-7asfkr7wt) · LangChain · 0 upvotes · 0 comments --- Tags: [#postgresql](https://daily.dev/tags/postgresql), [#rag](https://daily.dev/tags/rag), [#vector-search](https://daily.dev/tags/vector-search), [#full-text-search](https://daily.dev/tags/full-text-search) [View this post on daily.dev](https://daily.dev/posts/commsync-runs-text-vector-and-hybrid-search-on-postgres-with-lakebase-search-sybz5is1d) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"CommSync Runs Text, Vector, and Hybrid Search on Postgres with Lakebase Search","url":"https://daily.dev/posts/commsync-runs-text-vector-and-hybrid-search-on-postgres-with-lakebase-search-sybz5is1d","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/commsync-runs-text-vector-and-hybrid-search-on-postgres-with-lakebase-search-sybz5is1d"},"datePublished":"2026-07-30T16:30:04.104Z","dateModified":"2026-09-14T07:52:45.326Z","description":"CommSync, a unified inbox platform for business communications, uses Neon's Lakebase Search extensions (lakebase_text and lakebase_vector) to power three...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/510ba68fd4f14a1b6e2805db6f9b923a?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/510ba68fd4f14a1b6e2805db6f9b923a?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Neon","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Neon","logo":"https://media.daily.dev/image/upload/s--65EDghou--/f_auto/v1724339118/logos/neontech","url":"https://daily.dev/sources/neontech"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/commsync-runs-text-vector-and-hybrid-search-on-postgres-with-lakebase-search-sybz5is1d","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"postgresql,rag,vector-search,full-text-search","timeRequired":"PT5M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Neon","item":"https://daily.dev/sources/neontech"},{"@type":"ListItem","position":3,"name":"CommSync Runs Text, Vector, and Hybrid Search on Postgres with Lakebase Search"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/commsync-runs-text-vector-and-hybrid-search-on-postgres-with-lakebase-search-sybz5is1d#faq","mainEntity":[{"@type":"Question","name":"What is lakebase_text and how does it improve full-text search compared to native Postgres GIN indexes?","acceptedAnswer":{"@type":"Answer","text":"lakebase_text is a Postgres extension bundled in Neon's Lakebase Search that adds a lakebase_bm25 index type for full-text search. It works with standard tsvector types and operators but adds true BM25 relevance scoring and top-K pushdown, features native GIN indexes lack. One customer reported query latency dropping from about 19.5 seconds cold-start with GIN/tsquery to about 18.6ms warm. daily.dev surfaces real-world benchmarks like this for teams deciding between Postgres-native and extended full-text search."}},{"@type":"Question","name":"Is lakebase_vector compatible with existing pgvector queries?","acceptedAnswer":{"@type":"Answer","text":"Yes, lakebase_vector is a drop-in replacement for pgvector, using the same types, distance operators, and query syntax, so existing pgvector queries work unchanged. It adds the lakebase_ann index type for approximate nearest-neighbor search, while exact nearest-neighbor search with no index remains an option for small datasets needing perfect recall. Track Postgres vector search extensions like this on daily.dev before migrating a pgvector-based RAG pipeline."}},{"@type":"Question","name":"How can I combine keyword and semantic search into a single hybrid search ranking?","acceptedAnswer":{"@type":"Answer","text":"One production approach uses Reciprocal Rank Fusion (RRF) with the formula score = sum of 1/(k + rank) across a lexical arm and a semantic arm, using k=60. Weighting the lexical arm at 1.0 and the semantic arm at 0.8 lets exact keyword matches win ties over conceptually similar but non-literal semantic hits, useful when keyword traffic dominates. daily.dev helps developers comparing hybrid search ranking strategies keep up with real implementation details."}}]} ```