pg-aiguide

Name: pg-aiguide
Author: Tiger Data

Searches versioned PostgreSQL, TimescaleDB and PostGIS manuals and serves Postgres best-practice guides

Official · Tiger Data TypeScript Remote (hosted) · Streamable HTTP Local · stdio Databases Developer tools

GitHub npm Website

72 popularity score ★ 1.9k ↓ 162/wk

Connect your client

Remote (hosted)

Add in one click

Add to Cursor Add to LM Studio Add to Goose

Opens the app, which asks before it adds the server.

Or add it by command or config

Codexterminal

codex mcp add --url "https://mcp.tigerdata.com/docs" pg-aiguide

Cursor.cursor/mcp.json

{
  "mcpServers": {
    "pg-aiguide": {
      "url": "https://mcp.tigerdata.com/docs"
    }
  }
}

VS Codeterminal

code --add-mcp '{"name":"pg-aiguide","type":"http","url":"https://mcp.tigerdata.com/docs"}'

Claude CodeTerminal

claude mcp add --transport http pg-aiguide https://mcp.tigerdata.com/docs

Gemini CLIterminal

gemini mcp add -s user pg-aiguide "https://mcp.tigerdata.com/docs" -t http

opencode~/.config/opencode/opencode.json or a project opencode.json

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "pg-aiguide": {
      "type": "remote",
      "url": "https://mcp.tigerdata.com/docs"
    }
  }
}

windsurf~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "pg-aiguide": {
      "serverUrl": "https://mcp.tigerdata.com/docs"
    }
  }
}

Local (npm)

Add in one click

Add to Cursor Add to VS Code Add to LM Studio Add to Goose

Opens the app, which asks before it adds the server. Then replace each <NAME> in its settings with your value.

Or add it by command or config

It needs PGHOST, PGDATABASE, PGUSER, PGPASSWORD, OPENAI_API_KEY. Replace each <NAME> with your value.

Claude CodeTerminal

claude mcp add pg-aiguide -e PGHOST=<PGHOST> -e PGDATABASE=<PGDATABASE> -e PGUSER=<PGUSER> -e PGPASSWORD=<PGPASSWORD> -e OPENAI_API_KEY=<OPENAI_API_KEY> -- npx -y @tigerdata/pg-aiguide stdio

Claude Desktopclaude_desktop_config.json

{
  "mcpServers": {
    "pg-aiguide": {
      "command": "npx",
      "args": [
        "-y",
        "@tigerdata/pg-aiguide",
        "stdio"
      ],
      "env": {
        "PGHOST": "<PGHOST>",
        "PGDATABASE": "<PGDATABASE>",
        "PGUSER": "<PGUSER>",
        "PGPASSWORD": "<PGPASSWORD>",
        "OPENAI_API_KEY": "<OPENAI_API_KEY>"
      }
    }
  }
}

Cursor~/.cursor/mcp.json

{
  "mcpServers": {
    "pg-aiguide": {
      "command": "npx",
      "args": [
        "-y",
        "@tigerdata/pg-aiguide",
        "stdio"
      ],
      "env": {
        "PGHOST": "<PGHOST>",
        "PGDATABASE": "<PGDATABASE>",
        "PGUSER": "<PGUSER>",
        "PGPASSWORD": "<PGPASSWORD>",
        "OPENAI_API_KEY": "<OPENAI_API_KEY>"
      }
    }
  }
}

VS Code.vscode/mcp.json

{
  "servers": {
    "pg-aiguide": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@tigerdata/pg-aiguide",
        "stdio"
      ],
      "env": {
        "PGHOST": "<PGHOST>",
        "PGDATABASE": "<PGDATABASE>",
        "PGUSER": "<PGUSER>",
        "PGPASSWORD": "<PGPASSWORD>",
        "OPENAI_API_KEY": "<OPENAI_API_KEY>"
      }
    }
  }
}

Codex~/.codex/config.toml

[mcp_servers.pg-aiguide]
command = "npx"
args = ["-y", "@tigerdata/pg-aiguide", "stdio"]

[mcp_servers.pg-aiguide.env]
PGHOST = "<PGHOST>"
PGDATABASE = "<PGDATABASE>"
PGUSER = "<PGUSER>"
PGPASSWORD = "<PGPASSWORD>"
OPENAI_API_KEY = "<OPENAI_API_KEY>"

Gemini CLITerminal

gemini mcp add --scope user -e PGHOST=<PGHOST> -e PGDATABASE=<PGDATABASE> -e PGUSER=<PGUSER> -e PGPASSWORD=<PGPASSWORD> -e OPENAI_API_KEY=<OPENAI_API_KEY> pg-aiguide npx -- -y @tigerdata/pg-aiguide stdio

Local (from source)

Add in one click

Add to Cursor Add to VS Code Add to LM Studio

Opens the app, which asks before it adds the server. Then replace each <NAME> in its settings with your value.

Or add it by command or config

It needs PGHOST, PGDATABASE, PGUSER, PGPASSWORD, OPENAI_API_KEY. Replace each <NAME> with your value.

FirstTerminal

git clone https://github.com/timescale/pg-aiguide.git
cd pg-aiguide
./bun i

Claude CodeTerminal

claude mcp add pg-aiguide -e PGHOST=<PGHOST> -e PGDATABASE=<PGDATABASE> -e PGUSER=<PGUSER> -e PGPASSWORD=<PGPASSWORD> -e OPENAI_API_KEY=<OPENAI_API_KEY> -- node '<ABSOLUTE_PATH_TO_CLONE>/dist/index.js' stdio

Claude Desktopclaude_desktop_config.json

{
  "mcpServers": {
    "pg-aiguide": {
      "command": "node",
      "args": [
        "<ABSOLUTE_PATH_TO_CLONE>/dist/index.js",
        "stdio"
      ],
      "env": {
        "PGHOST": "<PGHOST>",
        "PGDATABASE": "<PGDATABASE>",
        "PGUSER": "<PGUSER>",
        "PGPASSWORD": "<PGPASSWORD>",
        "OPENAI_API_KEY": "<OPENAI_API_KEY>"
      }
    }
  }
}

Cursor~/.cursor/mcp.json

{
  "mcpServers": {
    "pg-aiguide": {
      "command": "node",
      "args": [
        "<ABSOLUTE_PATH_TO_CLONE>/dist/index.js",
        "stdio"
      ],
      "env": {
        "PGHOST": "<PGHOST>",
        "PGDATABASE": "<PGDATABASE>",
        "PGUSER": "<PGUSER>",
        "PGPASSWORD": "<PGPASSWORD>",
        "OPENAI_API_KEY": "<OPENAI_API_KEY>"
      }
    }
  }
}

VS Code.vscode/mcp.json

{
  "servers": {
    "pg-aiguide": {
      "type": "stdio",
      "command": "node",
      "args": [
        "<ABSOLUTE_PATH_TO_CLONE>/dist/index.js",
        "stdio"
      ],
      "env": {
        "PGHOST": "<PGHOST>",
        "PGDATABASE": "<PGDATABASE>",
        "PGUSER": "<PGUSER>",
        "PGPASSWORD": "<PGPASSWORD>",
        "OPENAI_API_KEY": "<OPENAI_API_KEY>"
      }
    }
  }
}

Codex~/.codex/config.toml

[mcp_servers.pg-aiguide]
command = "node"
args = ["<ABSOLUTE_PATH_TO_CLONE>/dist/index.js", "stdio"]

[mcp_servers.pg-aiguide.env]
PGHOST = "<PGHOST>"
PGDATABASE = "<PGDATABASE>"
PGUSER = "<PGUSER>"
PGPASSWORD = "<PGPASSWORD>"
OPENAI_API_KEY = "<OPENAI_API_KEY>"

Gemini CLITerminal

gemini mcp add --scope user -e PGHOST=<PGHOST> -e PGDATABASE=<PGDATABASE> -e PGUSER=<PGUSER> -e PGPASSWORD=<PGPASSWORD> -e OPENAI_API_KEY=<OPENAI_API_KEY> pg-aiguide node -- '<ABSOLUTE_PATH_TO_CLONE>/dist/index.js' stdio

Settings 8

`PGHOST`	Required: Host of the pgvector database holding the loaded docs (self-hosted only).
`PGDATABASE`	Required: Name of that database.
`PGUSER`	Required: Database user; a read-only user is enough.
`PGPASSWORD`	Required: Password for that user.
`OPENAI_API_KEY`	Required: OpenAI key used to embed search queries.
`PGPORT`	Optional: Database port.
`DB_SCHEMA`	Optional: Schema holding the docs tables; defaults to docs.
`DISABLE_MCP_SKILLS`	Optional: Set to 1 or true to remove view_skill and the skill prompts.

What it does

pg-aiguide is Tiger Data's (formerly Timescale) documentation server for coding assistants that write PostgreSQL. It has two tools. search_docs looks up passages in one documentation set at a time: the PostgreSQL manual for versions 14 to 18, PostGIS 3.3 to 3.6, or Tiger Data's own TimescaleDB and Tiger Cloud docs. It can match by keyword (BM25), by meaning (vector similarity), or blend the two. view_skill returns curated guidance documents on topics such as schema design, indexes, data types, constraints, naming, performance, exploring an existing schema and setting up TimescaleDB hypertables.

Most people use the free hosted endpoint at https://mcp.tigerdata.com/docs, which Tiger Data runs. You can also run it yourself from the npm package or a clone, but then you need your own pgvector database with the docs loaded into it, plus an OpenAI key for query embeddings. The same repository also ships the guides as agent skills and as a Claude Code plugin, so you can use them without the MCP server.

Some things the README leaves out, checked in the v0.6.1 source: search_docs takes a single source value that names both the manual and its version (postgres_14 to postgres_18, postgis_3.3 to postgis_3.6, or tiger). It returns 20 matches by default, and the code sets no upper limit on limit. semanticWeight defaults to 0.7, so searches lean toward meaning-based matches. Self-hosted, queries are embedded with OpenAI's text-embedding-3-small model.

What it needs

To run it yourself	Node.js

Hosted: an MCP client that can connect to a remote HTTP server. No account or key is needed. To run it yourself: Node.js, a PostgreSQL database with the pgvector extension, and the documentation scraped, embedded and loaded into it by following the repository's ingest directions (the database starts out empty). Embeddings need an OpenAI API key. Building from source uses the repo's ./bun script.

Limits

It searches only the bundled documentation sets: the PostgreSQL manual for versions 14 to 18, PostGIS 3.3 to 3.6, and Tiger Data's TimescaleDB and Tiger Cloud docs. Other versions and other extensions are not covered, and the README lists pgvector docs as still to come. It never connects to your own database or runs SQL. view_skill disappears when skills are turned off (disable_mcp_skills=1 on the hosted URL, DISABLE_MCP_SKILLS=1 for stdio), and the API docs say it is off in the Claude Code plugin, which loads the skills directly. A self-hosted copy returns nothing useful until you have loaded the docs yourself.

Who it suits

It suits developers who want their assistant to check the right Postgres manual version, or TimescaleDB and PostGIS docs, before writing schemas and queries. The hosted endpoint needs no key and nothing installed, so that is the easy route. Self-hosting is a real project (your own pgvector database, a docs load, an OpenAI key), so it makes sense only if you have to keep queries off Tiger Data's servers. It only looks up documentation. To inspect or change a live database, pair it with a database server.

Tools 2

search_docs: Searches one documentation set (PostgreSQL 14-18, PostGIS 3.3-3.6, or Tiger/TimescaleDB) by keyword, meaning or both
view_skill: Returns a curated Postgres or TimescaleDB best-practice guide by name, or a file inside it

Popularity

72 popularity score, on our list (the bar is 36)

★ 1.9k ↓ 162/wk

GitHub stars	1,884	39.3 points
Stars in 30 days	+25	7.1 points
npm downloads, last week	162	17.7 points
Official server	Yes	8 points

Measured 2026-11-07. How the score works.

Where it is published

Code	timescale/pg-aiguide ★ 1.9k
npm	@tigerdata/pg-aiguide 0.6.1, 2026-09-04
Hosted	`https://mcp.tigerdata.com/docs` Streamable HTTP · no sign-in

What it can reach

Runs	On your machine, or on the maker's servers
Reaches	The internet
Changes things	No: it reads only
Sign-in	None for the hosted endpoint; an API key or token to run it yourself
Licence	Apache-2.0 (npm)
Last code change	2026-11-06

Read-only: neither tool changes anything, and it never connects to your own databases. With the hosted endpoint, your search queries go to Tiger Data's server at mcp.tigerdata.com, which needs no key, and the code runs there. Run yourself, it connects with the PG* settings to the docs database you loaded, and it sends each meaning-based or blended query to OpenAI for embedding using your OPENAI_API_KEY. DB_SCHEMA is checked against a strict identifier pattern before it is used in SQL.

Over time

GitHub stars 1.9k +25 (+1 %)

1.9k1.9k

8 Oct 20267 Nov 2026

npm downloads a week 162 −209 (−56 %)

37113

15 Dec 20257 Nov 2026

Before the site measured it, read from npm's daily download counts.

Popularity score 72 +18 (+34 %)

7254

8 Oct 20267 Nov 2026

What is rising across the directory

Sources

Page checked 2026-10-15.

Badge for your README

Shows this server's rank in its category, and stars for a server that runs on your machine. It updates every hour.

MarkdownREADME.md

[![pg-aiguide on EveryMCPServer](https://everymcpserver.com/badge/pg-aiguide.svg)](https://everymcpserver.com/servers/pg-aiguide)

Connect your client

Add in one click

Or add it by command or config

Add in one click

Or add it by command or config

Add in one click

Or add it by command or config

Settings 8

What it does

What it needs

Limits

Who it suits

Tools 2

Similar servers

MCP Toolbox for Databases

PostgreSQL MCP Server

SocratiCode

Context

GitHits

Instant MCP