# Introducing pgContext 0.2.0: Advanced AI search inside Postgres
**作者**: dale
**日期**: 2026-07-23T14:47:33.000Z
**来源**: [https://x.com/daleverett/status/2080303809864331544](https://x.com/daleverett/status/2080303809864331544)
---

So. pgContext's launch went crazy viral 48 hours ago. But we weren't satisfied with just being 5.7x faster than pgVector and eliminating the need for indexes.
> TLDR: pgContext 0.2.0 makes advanced AI search inside PostgreSQL faster, broader, and easier to operate. Read below for more info
## Meet pgContext 0.2.0.
But first, if you're new here: pgContext is an open-source PostgreSQL extension for semantic search, filtering, hybrid retrieval, recommendations, and discovery. It gives Postgres vector superpowers. This was the original launch post if you'd like to understand how pgContext works.
It is available under the Apache 2.0 license for PostgreSQL 17.
[here’s the repo →]
Host on Polygres (free for limited time) → [Click Here]
## Updates at a glance
- More types of search can now be fast. Some advanced searches that previously had to scan everything no longer need to.
- You can see what the search system is doing. Teams get better visibility into speed, accuracy, and failures.
- Upgrades are safer. The release adds clearer handling for existing pgContext and pgvector setups.
- It supports more advanced AI search methods. Useful for teams building better recommendations, hybrid search, or lower-cost retrieval.
## Technical updates
pgContext 0.2.0 includes:
- Dense vector search with exact and HNSW-powered retrieval
- Metadata filtering across PostgreSQL columns and JSONB
- Hybrid search using vectors and PostgreSQL full-text search
- Collections over your existing tables, without copying your data
- Recommendations, discovery, facets, grouping, and scrolling
- Support for dense, sparse, half, bit, and multi-vector experiments
- SQL-based diagnostics for recall, index health, memory, and query performance
## Get started
Run pgContext yourself with the source release, container image, or Docker Compose playground.
For a hosted setup, you can sign up for free on Polygres.
Read the full release notes on GitHub.
GitHub → [Click Here]
Host on Polygres → [Click Here] (Free for limited time)
More links in the comments below :D
## 相关链接
- [dale](https://x.com/daleverett)
- [@daleverett](https://x.com/daleverett)
- [3.8K](https://x.com/daleverett/status/2080303809864331544/analytics)
- [[here’s the repo →]](https://github.com/evokoa/pgContext)
- [Host on Polygres (free for limited time) → [Click Here]](https://polygres.com/)
- [[Click Here]](https://github.com/evokoa/pgContext)
- [[Click Here]](https://polygres.com/)
- [Upgrade to Premium](https://x.com/i/premium_sign_up)
- [10:47 PM · Jul 23, 2026](https://x.com/daleverett/status/2080303809864331544)
- [3,822 Views](https://x.com/daleverett/status/2080303809864331544/analytics)
- [View quotes](https://x.com/daleverett/status/2080303809864331544/quotes)
---
*导出时间: 2026/7/24 10:20:37*
---
## 中文翻译
# 介绍 pgContext 0.2.0:Postgres 内部的高级 AI 搜索
**作者**: dale
**日期**: 2026-07-23T14:47:33.000Z
**来源**: [https://x.com/daleverett/status/2080303809864331544](https://x.com/daleverett/status/2080303809864331544)
---

所以。pgContext 在 48 小时前发布后迅速走红。但我们不仅仅满足于比 pgVector 快 5.7 倍并消除了对索引的需求。
> 简而言之:pgContext 0.2.0 让 PostgreSQL 内部的高级 AI 搜索变得更快、更广泛、更易于操作。阅读下文了解更多信息
## 认识 pgContext 0.2.0。
但首先,如果您是新来的:pgContext 是一个用于语义搜索、过滤、混合检索、推荐和发现的开源 PostgreSQL 扩展。它赋予了 Postgres 向量超能力。如果您想了解 pgContext 的工作原理,这是最初的发布帖。
它适用于 PostgreSQL 17,采用 Apache 2.0 许可证。
[这里是仓库 →]
在 Polygres 上托管(限时免费)→ [点击这里]
## 更新一览
- 更多种类的搜索现在可以变快。一些以前必须扫描所有内容的高级搜索不再需要这样做。
- 您可以看到搜索系统在做什么。团队可以更好地了解速度、准确性和故障。
- 升级更安全。该版本为现有的 pgContext 和 pgvector 设置添加了更清晰的处理。
- 它支持更高级的 AI 搜索方法。对于构建更好的推荐、混合搜索或低成本检索的团队非常有用。
## 技术更新
pgContext 0.2.0 包括:
- 使用精确和 HNSW 驱动检索的密集向量搜索
- 跨 PostgreSQL 列和 JSONB 的元数据过滤
- 使用向量和 PostgreSQL 全文搜索的混合搜索
- 在现有表上的集合,无需复制数据
- 推荐、发现、分面、分组和滚动
- 支持密集、稀疏、半精度、位和多向量实验
- 基于 SQL 的召回率、索引健康状况、内存和查询性能诊断
## 开始使用
使用源代码版本、容器镜像或 Docker Compose 游乐场自己运行 pgContext。
对于托管设置,您可以在 Polygres 上免费注册。
在 GitHub 上阅读完整的发布说明。
GitHub → [点击这里]
在 Polygres 上托管 → [点击这里](限时免费)
下面的评论中有更多链接 :D
## 相关链接
- [dale](https://x.com/daleverett)
- [@daleverett](https://x.com/daleverett)
- [3.8K](https://x.com/daleverett/status/2080303809864331544/analytics)
- [[here’s the repo →]](https://github.com/evokoa/pgContext)
- [Host on Polygres (free for limited time) → [Click Here]](https://polygres.com/)
- [[Click Here]](https://github.com/evokoa/pgContext)
- [[Click Here]](https://polygres.com/)
- [Upgrade to Premium](https://x.com/i/premium_sign_up)
- [10:47 PM · Jul 23, 2026](https://x.com/daleverett/status/2080303809864331544)
- [3,822 Views](https://x.com/daleverett/status/2080303809864331544/analytics)
- [View quotes](https://x.com/daleverett/status/2080303809864331544/quotes)
---
*导出时间: 2026/7/24 10:20:37*