See what's new.
The analytics backend for your app and agents
Tinybird is the real-time analytics platform for the agentic era. Learn more about the Tinybird MCP Server, the Explorations UI, and how to build your own autonomous analytics agents.
[01]
Tinybird MCP
Your real‑time data, LLM‑ready
A remote, hosted MCP server that allows your AI agents to connect directly to your Tinybird workspace.
Your Tinybird workspaces as remote MCPs
Instantly connect LLMs and agents to any Tinybird workspace. Tinybird's fully managed, hosted MCP server gives you everything you need to expose your data and APIs as LLM tools—no setup, no ops, zero friction.
// Get your TB_TOKEN from https://cloud.tinybird.co/tokens
// Configure your client with the Tinybird MCP server URL and a token
{
"mcpServers":{
"tinybird":{
"url":"https://mcp.tinybird.co?token=TB_TOKEN"
}
}
}Advanced Data Tools, Out‑of‑the‑Box
Tinybird's hosted MCP server exposes convenient tools for agents to successfully and efficiently explore Tinybird data
from agno.agent import Agent
from agno.models.anthropic import Claude
from agno.tools.mcp import MCPTools
import asyncio
import os
tinybird_api_key = os.getenv("TINYBIRD_TOKEN")
tinybird_host = os.getenv("TINYBIRD_HOST")
async def main():
async with MCPTools(
transport="streamable-http",
url=f"https://mcp.tinybird.co?token={tinybird_api_key}&host={tinybird_host}",
timeout_seconds=120) as mcp_tools:
agent = Agent(model=Claude(id="claude-4-opus-20250514"), tools=[mcp_tools])
await agent.aprint_response("top 5 pages with more visits in the last 24 hours", stream=True)
if __name__ == "__main__":
asyncio.run(main())No Data Leakage
MCP never exposes more than the underlying credentials allow.
/ 01
Access Control Built In
Tinybird MCP leverages your existing Tinybird static tokens and JWTs.
/ 02
RBAC for your MCPs
Give precise access with row-level authorization.
Built in observability
Every tool call is tracked and stored in Tinybird service data sources, so you have detailed observability about how AI is using your resources.
Works with Cursor, clients, and SDKs
Any agent, IDE, or client SDK can interact with Tinybird via Streamable HTTP.
[02]
Birdwatcher Analytical Agents and Templates
Analytics agents that you can install or clone to fit your purposes
Birdwatcher Notifications
Give this background agent a mission, and it will monitor your Tinybird workspace according to the mission and notify you as instructed.
Health templates
Includes default prompts to monitor the health of your Tinybird cluster.
Slack or Email notifications
Choose to be notified via Slack, Email, or both. Run it on GitHub Actions, as a cron job, or on demand.
Open Source
Clone the repo and build your own analytics agent in a few minutes.
[See the Birdwatcher repo]# Clone the repo
git clone git@github.com:tinybirdco/ai.git
cd ai/agents/birdwatcher
# Fill in the required environment variables for your Tinybird account and LLMs.
# Optionally, provide your Slack or Resend API keys to enable notifications
cp .env.example .env
# Install uv
curl -LsSf https://astral.sh/uv/install.sh | sh
uv sync
# Run the agent
uv run python birdwatcher.py \
--prompt "analyse website visits and notify me on #tmp-birdwatcher" \
--mission baseBirdwatcher Slack
A Slack bot to talk to your Tinybird data. Birdwatcher will answer any question about your Tinybird data or usage.
Add to Slack
Install Birdwatcher
Includes default prompts to monitor the health of your Tinybird cluster.
Invite to your channel or DM
If you are a Tinybird customer, simply add it to your Slack workspace and configure your tokens.
Open Source
Clone the repo, define a new mission, and deploy it. It's that simple.
[See the Birdwatcher repo][03]
Explorations
A natural language exploratory data analysis UI to help you understand the shape of your data.
Query your data with natural language
Dig into your data without knowing its schema. Tinybird has all the context to write SQL that answers your questions.
Explore iteratively
Tinybird retains the context from your chat, so you can perform exploratory data analysis interactively and iteratively.
Auto-generated results, analysis and charts
When prompted, Explorations generates timeseries visualizations to help you understand how your data is evolving.
Tinybird is the data platform for real‑time agentic analytics.
