v2.3.1 Release Notes

UX Improvements

Refreshed publish note UI with new brand colors, improved readability, etc.
Improved speaker identification UX. Refined the interface for identifying and labeling speakers in meeting transcripts
Updated project experience with a rearchitected data layer for better reliability and performance
Removed audio import length cap for free users. The 45-minute limit on audio file imports has been lifted for free accounts.
The AI agent is better at searching for the right tools to use

MCP tools that use dynamic field definitions were sometimes incorrectly sending empty objects; this is now resolved
Added handling for missing file errors that were surfacing as unhandled exceptions