Tinybird Code improvements and more¶

Highlights

• Tinybird Code now handles large projects more efficiently with improved context management
• Viewers and guests can validate deployments with tb --cloud deploy --check
• Faster query processing and improved Endpoint response times
• Enhanced UI stability with fixes for browser compatibility and modal behavior
• Better error messages and improved Token management visibility

Tinybird Code¶

We've optimized the Tinybird Code agent framework to handle complex, multi-turn conversations without losing critical context, while also making efficient use of the context window.

TL;DR

• More reliable performance with larger projects
• Better code suggestions thanks to improved project understanding
• Faster responses by only loading what's needed

Requires upgrading to tinybird 1.1.10

Improved Context Management

We've optimized how Tinybird Code handles large projects. Previously, when working with projects containing many files or large datafiles, the system could run into context limits too quickly. Now, Tinybird Code loads file content on-demand, only fetching what's needed. This means you can work with larger, more complex projects without hitting limits prematurely.

Enhanced Project Understanding

The system now better understands your project structure by tracking dependencies between files (like which pipes use which Data Sources) without needing to read every file upfront. This helps Tinybird Code provide more accurate suggestions and understand your codebase more efficiently.

Model Upgrade

We've upgraded to a newer, more capable AI model that provides better code suggestions and understanding of your Tinybird projects.

Improvements and bug fixes¶

Performance improvements¶

• Faster Query Processing: Multiple optimizations to memory usage and database operations resulting in more efficient query execution
• Improved Response Times: Endpoints now serve responses faster with updated latency thresholds
• Enhanced Import Processing: Increased NDJSON import processing capacity for large datasets

User interface enhancements¶

• Fixed issues with the combobox display to prevent text truncation
• Resolved browser compatibility problems in Firefox
• Fixed UI crashes when using Google Translate
• Improved modal behavior to prevent unexpected closures

New features & capabilities¶

• Expanded Permissions: Viewers and guests can now use the deploy --check command to validate deployments without making changes
• Better Token Management: Improved token visibility in the UI, only showing tokens users have permission to view or modify

Improved error handling¶

• More helpful error messages for credential issues
• Better feedback when sink pipe query parameter secrets are missing
• Improved error messaging for connection problems

Bug fixes¶

• Deployment Fixes:
  • Fixed issues with secrets handling in base pipes for sinks
  • Ensured proper database creation sequence during deployment
  • Fixed branch workspace organization alignment
• Critical Fixes:
  • Resolved circular dependency issues
  • Fixed authentication edge cases with JWT handling