pg-aiguide

pg-aiguide is an MCP server that connects AI coding assistants to version-aware PostgreSQL documentation, ecosystem manuals, and curated database engineering skills. Developed by Timescale, the server is designed for software engineers, database architects, and application developers who rely on coding agents like Cursor, Claude Code, Windsurf, VS Code, and Codex to design schemas and write SQL. AI tools often generate database definitions that lack constraints, rely on deprecated syntax, or miss opportunities for modern index optimization. To address this, pg-aiguide provides specialized tools that perform semantic vector similarity and keyword BM25 search across official PostgreSQL documentation versions, PostGIS manuals, and TimescaleDB references. Additionally, it provides an AI-optimized skill registry containing opinionated, battle-tested guidelines for schema design, constraints, indexing strategies, naming conventions, and performance tuning. By referencing actual manual excerpts and modern PostgreSQL features such as generated identity columns and distinct null handling, pg-aiguide enables AI assistants to draft robust, maintainable, and high-performance database architectures directly within the developer workflow.

Category: Databases & Data Stores

Tags: coding-assistant, documentation, postgres, SQL

Visit pg-aiguide

How to install and configure pg-aiguide

1. Open your MCP client configuration file, such as .cursor/mcp.json in Cursor or ~/.codeium/windsurf/mcp_config.json in Windsurf. 2. Add the remote HTTP server configuration pointing to the TigerData endpoint: json { "mcpServers": { "pg-aiguide": { "url": "https://mcp.tigerdata.com/docs" } } } 3. If using VS Code from the command line, register the server by executing: code --add-mcp '{"name":"pg-aiguide","type":"http","url":"https://mcp.tigerdata.com/docs"}' 4. If using Claude Code, install the official plugin by running claude plugin marketplace add timescale/pg-aiguide and then claude plugin install pg@aiguide. 5. Restart your AI client to enable the search_docs and view_skill tools.

What you can do with pg-aiguide

• Generating optimized PostgreSQL schemas equipped with modern constraints, identity columns, and domain-appropriate primary keys using curated skill templates. * Searching version-specific PostgreSQL documentation via semantic or keyword queries to verify syntax and configuration settings for exact engine versions. * Retrieving specialized documentation and best practices for extensions like TimescaleDB hypertables and PostGIS spatial indexing directly into agent conversations. * Reviewing existing database schema definitions to identify missing indexes, deficient check constraints, or suboptimal data types based on curated guidelines. * Designing time-series or IoT data ingestion schemas that require composite indexing, partitioning rules, and advanced historical querying configurations.

Key facts

• https://github.com/timescale/pg-aiguide
• Databases & Data Stores, Developer Tools & Code Intelligence
• coding-assistant, documentation, postgres, SQL

Part of MCP Servers

What is pg-aiguide?

pg-aiguide is an MCP server and agent skill suite developed by Timescale that provides AI coding tools with direct access to official PostgreSQL documentation and curated engineering skills. It helps coding assistants write accurate, modern, and performant SQL by providing semantic search across documentation and opinionated best practices for schema design and indexing.

What can pg-aiguide do?

pg-aiguide exposes tools like search_docs and view_skill to AI assistants. The search tool performs semantic and keyword BM25 queries across official PostgreSQL manuals, TimescaleDB documentation, and PostGIS guides. The skill tool delivers curated instructions covering schema architecture, constraints, indexing strategies, naming conventions, and performance optimization.

Which MCP clients work with pg-aiguide?

pg-aiguide works with any MCP-compatible environment that supports remote HTTP server connections or agent skills. Supported tools include Cursor, Claude Code, VS Code, VS Code Insiders, Visual Studio, Windsurf, Codex, Gemini CLI, OpenCode, Goose, and LM Studio.

How do I install pg-aiguide?

You can connect your client to the public MCP endpoint at https://mcp.tigerdata.com/docs by adding it to your client configuration file. In Claude Code, you can install the plugin directly from the marketplace using the claude plugin commands, or install agent skills using npx skills add timescale/pg-aiguide.

What documentation sources does pg-aiguide support?

pg-aiguide includes version-specific documentation from the official PostgreSQL manual, TimescaleDB documentation from Tiger Data, and spatial extension references from PostGIS. The system allows AI agents to scope queries to specific documentation subsets and engine releases to ensure relevant advice.