io.github.wenRou21/sqlbot-mcp-proxy

io.github.wenRou21/sqlbot-mcp-proxy

Enables natural-language querying of databases via SQLBot, with datasource discovery, metadata inspection, and report generation.

Category
Visit Server

README

SQLBot MCP Proxy

MCP server proxy for SQLBot. It exposes SQLBot datasource discovery, table metadata, natural-language Text-to-SQL questions, and local fallback report generation as MCP tools.

<!-- mcp-name: io.github.wenRou21/sqlbot-mcp-proxy -->

Requirements

  • Docker
  • A running SQLBot service
  • SQLBot API keys

Recommended Codex Config

Users do not need to download this repository manually. Add this single MCP server to Codex config and replace the SQLBot URL and keys:

[mcp_servers.sqlbot]
command = "docker"
args = [
  "run",
  "--rm",
  "-i",
  "--add-host",
  "host.docker.internal:host-gateway",
  "-e",
  "SQLBOT_MCP_PROXY_BASE_URL",
  "-e",
  "SQLBOT_MCP_PROXY_ACCESS_KEY",
  "-e",
  "SQLBOT_MCP_PROXY_SECRET_KEY",
  "ghcr.io/wenrou21/sqlbot-mcp-proxy:0.1.0"
]
enabled = true
startup_timeout_sec = 120

[mcp_servers.sqlbot.env]
SQLBOT_MCP_PROXY_BASE_URL = "http://host.docker.internal:8000"
SQLBOT_MCP_PROXY_ACCESS_KEY = "<YOUR_SQLBOT_ACCESS_KEY>"
SQLBOT_MCP_PROXY_SECRET_KEY = "<YOUR_SQLBOT_SECRET_KEY>"

Then restart Codex completely.

If SQLBot is reachable through a public or LAN address, replace SQLBOT_MCP_PROXY_BASE_URL with that address, for example https://sqlbot.example.com or http://192.168.1.10:8000.

Prompt Template

Please configure SQLBot MCP for Codex.

MCP image:
ghcr.io/wenrou21/sqlbot-mcp-proxy:0.1.0

My SQLBot URL:
http://host.docker.internal:8000

My SQLBot access key:
<YOUR_SQLBOT_ACCESS_KEY>

My SQLBot secret key:
<YOUR_SQLBOT_SECRET_KEY>

Add this MCP to Codex config:
- MCP name: sqlbot
- command: docker
- args: ["run", "--rm", "-i", "--add-host", "host.docker.internal:host-gateway", "-e", "SQLBOT_MCP_PROXY_BASE_URL", "-e", "SQLBOT_MCP_PROXY_ACCESS_KEY", "-e", "SQLBOT_MCP_PROXY_SECRET_KEY", "ghcr.io/wenrou21/sqlbot-mcp-proxy:0.1.0"]
- startup_timeout_sec: 120
- SQLBOT_MCP_PROXY_BASE_URL: use the SQLBot URL above
- SQLBOT_MCP_PROXY_ACCESS_KEY: use the access key above
- SQLBOT_MCP_PROXY_SECRET_KEY: use the secret key above

After configuration, remind me to fully restart Codex.

After restart:
- Check available datasources with sqlbot_list_datasources.
- Use sqlbot_list_tables and sqlbot_describe_table before asking data questions.
- Use sqlbot_ask_data for Text-to-SQL questions.
- Use sqlbot_generate_report when I ask for an analysis report or chart.

Quick Docker Test

You can test the image outside Codex:

docker run --rm -i \
  --add-host host.docker.internal:host-gateway \
  -e SQLBOT_MCP_PROXY_BASE_URL=http://host.docker.internal:8000 \
  -e SQLBOT_MCP_PROXY_ACCESS_KEY=your-access-key \
  -e SQLBOT_MCP_PROXY_SECRET_KEY=your-secret-key \
  ghcr.io/wenrou21/sqlbot-mcp-proxy:0.1.0

Tools

  • sqlbot_list_datasources: list SQLBot datasources visible to the configured API key
  • sqlbot_list_tables: list checked SQLBot metadata tables for a datasource
  • sqlbot_describe_table: describe checked fields for a SQLBot metadata table
  • sqlbot_ask_data: ask SQLBot a natural-language data question
  • sqlbot_generate_report: generate an analysis report and chart, with a local SVG fallback

Environment Variables

Name Default Description
SQLBOT_MCP_PROXY_BASE_URL http://127.0.0.1:8000 SQLBot service base URL
SQLBOT_MCP_PROXY_ACCESS_KEY empty SQLBot API access key
SQLBOT_MCP_PROXY_SECRET_KEY empty SQLBot API secret key
FEISHU_ASK_ACCESS_KEY empty Fallback access key name
FEISHU_ASK_SECRET_KEY empty Fallback secret key name
SQLBOT_MCP_PROXY_DOCKER_CONTAINER sqlbot Container used for fallback key lookup
SQLBOT_MCP_PROXY_HTTP_HOST 127.0.0.1 HTTP mode host
SQLBOT_MCP_PROXY_HTTP_PORT 8787 HTTP mode port
SQLBOT_MCP_PROXY_HTTP_PATH /mcp HTTP mode JSON-RPC path
SQLBOT_MCP_PROXY_HTTP_TIMEOUT_MS 120000 SQLBot HTTP request timeout
SQLBOT_MCP_PROXY_REPORT_OUTPUT_DIR outputs Directory for generated fallback SVG reports

Advanced HTTP Mode

Most MCP clients should use stdio. For clients that can send JSON-RPC over HTTP, start:

docker run --rm -p 8787:8787 \
  --add-host host.docker.internal:host-gateway \
  -e SQLBOT_MCP_PROXY_BASE_URL=http://host.docker.internal:8000 \
  -e SQLBOT_MCP_PROXY_ACCESS_KEY=your-access-key \
  -e SQLBOT_MCP_PROXY_SECRET_KEY=your-secret-key \
  -e SQLBOT_MCP_PROXY_HTTP_HOST=0.0.0.0 \
  ghcr.io/wenrou21/sqlbot-mcp-proxy:0.1.0 --http

Health check:

curl http://127.0.0.1:8787/health

Publish

  1. Confirm package metadata in package.json.
  2. Publish an OCI image:
docker build -t ghcr.io/wenrou21/sqlbot-mcp-proxy:0.1.0 .
docker push ghcr.io/wenrou21/sqlbot-mcp-proxy:0.1.0
  1. Publish MCP Registry metadata:
mcp-publisher login github
mcp-publisher publish

The mcpName in package.json must exactly match name in server.json.

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured