Birdwatcher is an AI agent that talks to your data in Tinybird, automating data exploration, monitoring, and complex analysis.
It uses the Tinybird Remote MCP Server to give you access to advanced data tools for building analytics agents over your data.
Today, we’re excited to announce the Birdwatcher Slack App, a conversational agent that lets you explore your Tinybird data directly from Slack.
How to get started
To begin, install Birdwatcher in your Slack workspace and run /birdwatcher-config in any public or private channel.
Alternatively, open a direct message with @Birdwatcher and run /birdwatcher-config. This allows you to connect Birdwatcher to your Tinybird workspace in a private conversation.
If you’re a Tinybird client, start a private chat with @Birdwatcher in the Tinybird Slack workspace. No installation required.
Once configured, you can use natural language to query your data or request key organization metrics. Examples:
- "What's my storage usage?"
- "What are the pipes with the most errors in the last 24 hours?"
- "Which datasources have the most rows in quarantine in the last hour?"
How it works
Birdwatcher is a standard Slack app that communicates with a web server. This server routes Slack event subscriptions to the Birdwatcher agent, which handles the requests.
Architecture overview
┌─────────────────┐
│ Slack App │
│ │
└─────────────────┘
|
event subscription
|
v
┌─────────────────┐
│ Birdwatcher Web │
│ Server │
└─────────────────┘
|
v
┌─────────────────┐
│ Birdwatcher │
│ Agent │
└─────────────────┘
|
+--------+--------+
| |
get user/channel token token
| |
v v
┌────────────────────┐ ┌─────────────────┐
│ birdwatcher_config │ │ Tinybird │
│ │ │ MCP │
└────────────────────┘ └─────────────────┘
|
token
|
v
┌─────────────────┐
│ Your Tinybird │
│ Workspace │
└─────────────────┘
Each Slack channel has its own @Birdwatcher configuration. The /birdwatcher-config command is scoped to the channel it runs in. In DMs, the config is scoped to the user.
Hybrid mode: Notifications + Conversations
You can combine Birdwatcher Notifications (running in the background) with the conversational Slack agent to monitor and investigate issues in your Tinybird workspace.
For instance, you can deploy a Birdwatcher agent using GitHub Actions with a given mission (See daily_summary.md), in this case investigate CPU spikes in your Tinybird cluster:
name: Birdwatcher Agent
on:
schedule:
- cron: '0 * * * *' # Run hourly
jobs:
monitor-cpu-spikes:
runs-on: ubuntu-latest
steps:
- uses: tinybirdco/ai@main
with:
slack_token: ${{ secrets.SLACK_TOKEN }}
anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
tinybird_token: ${{ secrets.TINYBIRD_TOKEN }}
tinybird_host: ${{ secrets.TINYBIRD_HOST }}
prompt: |
Report daily metrics about pipe_stats, datasources_ops_log and
jobs_log in the #tmp-birdwatcher Slack channel. No markdown.
mission: daily_summary
model: claude-4-sonnet-20250514
Once you receive a Slack notification from the Birdwatcher agent, follow up with questions to @Birdwatcher to investigate the issue further in natural language.
@Birdwatcher uses the explore mission which leverages the explore_data tool in the Tinybird MCP server.
Learn more
Try the Birdwatcher quickstart to deploy or build your own analytical agent.
