May was one of our biggest release months yet.
We introduced Model Context Protocol (MCP) support, launched an Amazon Redshift connector, made scheduled imports far more reliable, improved AI Analyst, and redesigned how you inspect import runs.
Here's everything that shipped.
🤖 Turn your data sources into AI tools with MCP
One of our biggest launches this month was native Model Context Protocol (MCP) support.
Instead of exporting data into AI tools, Brooked lets AI connect directly to your live data sources.
Connect Snowflake, Amazon Redshift, PostgreSQL, QuickBooks, and more, then use them instantly from clients like Claude Desktop, Cursor, or any MCP-compatible application.
Highlights include:
- Secure API key management
- Automatic tool generation for every connected data source
- Full support for the latest MCP protocol
- Streaming responses for faster AI interactions
No custom integrations. No glue code.
📊 Amazon Redshift is now supported
Amazon Redshift joins our growing list of native connectors.
You can now:
- Query tables and views
- Run custom SQL
- Sync results directly into Google Sheets
- Schedule automatic refreshes
We also increased SQL connection timeouts across warehouse connectors, making large analytical queries much more reliable.
🔄 Scheduled imports now work even when you're offline
Scheduled jobs shouldn't fail just because your browser session expired.
Until now, Google OAuth could interrupt background imports if authentication expired.
We've fixed that by adding offline authentication support. Brooked now securely stores refresh tokens and renews access automatically, allowing scheduled imports to continue running without manual intervention.
We also improved error messages when destination spreadsheets are deleted, replacing cryptic API responses with clear, actionable guidance.
📈 AI Analyst can now generate Mermaid diagrams
Sometimes a SQL result isn't enough.
When explaining data models, pipelines, or entity relationships, AI Analyst can now generate Mermaid diagrams directly inside conversations.
Instead of describing a workflow in text, Brooked visualizes it automatically, making complex systems much easier to understand.
🔍 Better run history and debugging
We redesigned the run history experience.
Instead of opening a small modal, every import, export, and alert now opens in a dedicated slide-over panel with:
- Complete execution logs
- Row counts
- Timing information
- Detailed error messages
This makes troubleshooting significantly faster without leaving your current workflow.
What's next?
We're already working on the next set of improvements, including:
- Smarter auto-pause behavior for scheduled imports
- More native connectors
- Deeper AI Analyst integration with scheduled workflows
As always, we're shipping based on customer feedback.
Need help?
If you have questions, run into an issue, or want to request a new connector, visit our support center:
You'll find documentation, troubleshooting guides, and answers to common questions.
Thanks for following along—we've got plenty more coming.
Top comments (0)