Version 1.7.8 - Release notes

Sujith Aleshwaram

— 2 min read
Version 1.7.8 - Release notes

Version 1.7.8

We have just shipped an update to the Locai L1-Large model with some bug fixes and improvements (Version 1.7.8).

A huge thank you to the Locai community for sharing your feedback on what's working, what's not, and what you want to see from Locai moving forward.

We read all the feedback that comes in - please keep it coming. It lets us know how we're doing, helps us prioritise future features and ensures we keep pointing in the right direction.

Sending us feedback is super simple. When you're next chatting with Locai, click on your avatar and select "Send Feedback" and use the drop down to report a bug, tell us about a feature you'd like to see or just drop us a line about anything else (within reason).

You can also suggest features by jumping onto our Discord server.

WHAT'S NEW

BUG FIXES

  • Code Block Copy Button - the copy button wasn’t always playing ball when copying code snippets. That bug has now been squashed.
  • Export file authentication - we fixed an issue where exported file links sometimes showed as not authenticated, ensuring reliable file access.
  • Smoother mobile scrolling - we’ve improved auto-scroll cancel behaviour on mobile devices.
  • LaTeX Copy formatting - a fix to ensure cleaner LaTeX code copying.
  • Feedback form - we have fixed issues related to drag-and-drop file uploads and the stability of the user feedback pop-up window.
  • Conversation Context Loss - we have fixed an issue where the model would occasionally lose conversation context during extended chat sessions (even AI minds wander from time to time it seems), ensuring better context retention and continuity.
  • Founding Member Badge fix - Founding Member Badges are now appearing as they should across all screen sizes. Founding Members, we appreciate you!

KNOWN ISSUES (AKA “WE’RE WORKING ON IT”)

  • Response crash recovery - in rare cases, when Locai crashes while writing a response, some users may get a message reading “Thinking…” despite the response having timed out.
  • Agent context awareness - we are working on delivering more personalised responses by improving how we incorporate user-defined context from users’ account preferences.

Read more