{"slug": "pgsync-adds-semantic-search-without-data-leaving-your-infrastructure", "title": "PGSync adds semantic search without data leaving your infrastructure", "summary": "PGSync Pro adds real-time semantic and hybrid search to PostgreSQL, re-embedding vectors only when meaning changes and keeping data within the user's infrastructure. The tool, built on the open-source PGSync, uses CDC to maintain fresh vector indexes and supports local sentence-transformers embeddings with optional OpenAI, Cohere, or Voyage integrations. A live demo shows that searching \"a rebellion against a watchful state\" returns George Orwell's \"Nineteen Eighty-Four\" by matching meaning rather than keywords.", "body_md": "# PGSync Pro[¶](#pgsync-pro)\n\nReal-time **semantic & hybrid search** for PostgreSQL — that only re-embeds\nwhen *meaning* changes.\n\nYour database stays the source of truth — PGSync Pro keeps a fresh vector index\nin lock-step, and only pays to embed when *meaning* actually changes. Built on\nPGSync, which syncs **PostgreSQL, MySQL and MariaDB** to Elasticsearch /\nOpenSearch.\n\nSee it live — no signup\n\nType an *idea*, not keywords. On the live demo, searching **“a rebellion\nagainst a watchful state”** returns **Nineteen Eighty-Four** — a book that\nshares none of those words. That's the difference between matching *words* and\nmatching *meaning*, kept in sync with the database in real time.\n\n## What you get[¶](#what-you-get)\n\n-\n**Always fresh**\n\nVectors stay in sync with your database via CDC. No nightly reindex jobs, no stale search results.\n\n-\n**The change-guard**\n\nRe-embed only when the embedded text changes. Unrelated column updates are free — and easy on your embedding bill.\n\n-\n**Semantic + hybrid**\n\nVector kNN fused with keyword BM25 (reciprocal rank fusion) — the best of meaning and exact match.\n\n-\n**Runs in your infra**\n\nLocal\n\n`sentence-transformers`\n\nembeddings mean text never leaves your environment. OpenAI / Cohere / Voyage optional. -\n**Agent-ready (MCP)**\n\nAn MCP server exposes search to Claude Desktop, Claude Code, or your own agent — retrieval over live data.\n\n-\n**Preflight checks**\n\n`pgsync-pro validate`\n\nverifies your schema and connections before indexing — catching mistakes with actionable errors.\n\nPGSync and django-pgsync stay free and MIT forever. Pro adds the semantic layer\non top — the same `schema.json`\n\nand CLI, plus an `embedding`\n\nblock.\n\n**Next:** try the live demo, watch the pipeline animated, then pick a plan.", "url": "https://wpnews.pro/news/pgsync-adds-semantic-search-without-data-leaving-your-infrastructure", "canonical_source": "https://pgsync.com/pro/", "published_at": "2026-07-26 00:30:48+00:00", "updated_at": "2026-07-26 00:52:16.423895+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "natural-language-processing", "developer-tools", "ai-products"], "entities": ["PGSync Pro", "PostgreSQL", "Elasticsearch", "OpenSearch", "OpenAI", "Cohere", "Voyage", "Claude Desktop"], "alternates": {"html": "https://wpnews.pro/news/pgsync-adds-semantic-search-without-data-leaving-your-infrastructure", "markdown": "https://wpnews.pro/news/pgsync-adds-semantic-search-without-data-leaving-your-infrastructure.md", "text": "https://wpnews.pro/news/pgsync-adds-semantic-search-without-data-leaving-your-infrastructure.txt", "jsonld": "https://wpnews.pro/news/pgsync-adds-semantic-search-without-data-leaving-your-infrastructure.jsonld"}}