Tinybird MCP Server (Deprecated, Unmaintained Repository)

Overview

This repository contains a minimal Model Context Protocol (MCP) server implementation that was used to forward and adapt event data for Tinybird analytics. It provides a lightweight proxy/agent pattern for receiving events, applying simple routing or transformation rules, and forwarding them to downstream analytics endpoints.

Important: this repository is deprecated and no longer maintained. For production use or ongoing support, migrate to the official Tinybird MCP Server implementation and follow the maintained documentation and migration guidance: https://www.tinybird.co/docs/forward/analytics-agents/mcp

This article explains what the project does, how it was typically run, and common configuration and use cases to help developers understand the repository contents and plan migration.

Features

• Receives incoming event payloads (HTTP/webhook) and forwards them to configured backends
• Config-driven routing and simple transformation rules
• Pluggable authentication and filtering hooks
• Lightweight, suitable for running close to the source of events (edge, containers)
• Basic logging and metrics endpoints for health checks and debugging

Installation / Configuration

Note: the example commands below are generic — inspect the repository (package.json, Dockerfile, requirements.txt, etc.) to determine the concrete runtime and dependencies before running.

Clone the repository:

git clone https://github.com/tinybirdco/mcp-tinybird.git
cd mcp-tinybird

Common ways to run the server:

• With Docker (recommended when available)

# Build image (if Dockerfile present)
docker build -t mcp-tinybird .

# Run container with environment variables and mounted config
docker run -d --name mcp-tinybird \
  -p 8080:8080 \
  -v "$(pwd)/config:/app/config:ro" \
  -e PORT=8080 \
  -e TINYBIRD_API_TOKEN="your-token" \
  mcp-tinybird

• Locally (Node.js / Python – check repo files)

# If Node.js
npm install
npm start

# Or if Python and requirements.txt present
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
python main.py

Example environment variables and configuration file layout (typical keys)

Example config snippet (YAML)

routes:
  - name: page_views
    path: /collect/page
    method: POST
    forward_to: https://api.tinybird.co/v0/pipes/ingest_page/append
    headers:
      Authorization: "Bearer ${TINYBIRD_API_TOKEN}"
    transforms:
      - type: remove_fields
        fields: ["debug"]

Testing the server (generic curl example)

# Send an event to a configured route
curl -X POST http://localhost:8080/collect/page \
  -H "Content-Type: application/json" \
  -d '{"user_id":"u123","page":"/home","timestamp":"2026-04-10T12:00:00Z"}'

If the repository exposes health or metrics endpoints, confirm them:

curl http://localhost:8080/health
curl http://localhost:8080/metrics

Available Resources

• Deprecated repository (reference only): https://github.com/tinybirdco/mcp-tinybird
• Official, maintained implementation and documentation (migrate here): https://www.tinybird.co/docs/forward/analytics-agents/mcp
• Migration guidance: consult the official docs link above for recommended migration steps, configuration mapping, and supported features
• Issues & community: check the GitHub repository for legacy issues and discussions, but do not expect active maintenance

Use Cases

1. Collecting browser or app analytics and forwarding to Tinybird

   • Deploy the MCP server close to your front-end or CDN edge.
   • Accept JSON events on a lightweight endpoint, apply a transform to normalize fields, and forward batched data to Tinybird ingestion endpoints.

2. Webhook aggregation and routing

   • Use the server as a central ingestion point for multiple webhook sources.
   • Configure routes for different providers, enrich or filter events, and forward to dedicated Tinybird pipes or other analytics backends.

3. Edge preprocessing for privacy or bandwidth

   • Strip or redact PII fields (e.g., remove email or IP) via transform rules before forwarding.
   • Aggregate or sample events to reduce bandwidth to the central analytics service.

4. Temporary gateway during migration

   • Use the MCP server as a transitional adapter while migrating services to the official Tinybird agent.
   • Configure equivalent routes to mirror traffic patterns and validate parity before switching production traffic.

Notes and Migration Advice