Skip to content

Version 0.3.0

Choose a tag to compare

View all tags
@Tomobobo710 Tomobobo710 released this 03 Jul 15:48
· 120 commits to main since this release
v0.3.0
6e0360c

Release Notes

This release includes the following new features and improvements:

New Features

Edit

  • Any message in any turn (user or assistant) can be edited to directly modify the chat history
  • An edit modal will appear over top of the turn and allows the user to edit individual messages per turn
  • For this action, chat history is permanently edited and used in future queries

Edit and Retry

  • Any user turn can be edited and sequentially sends a new query at this point in the conversation
  • A navigation action will be available for the user to browse "branches" that are created at this point of divergence

Retry

  • Any assistant response can be retried to produce an alternative response to any user query
  • Similarly, navigation actions will be available for these branching points

Chat Branch Navigation

  • Any point of divergence will result in < and > navigation actions available in the action bar located below turn content
  • The user can navigate between branches which connect at this point
  • Only the active branch's connection points have navigation actions

Improved Reasoning Support

  • Models which use <think> or <thinking> XML tags in content will have their reasoning displayed in "thinking" dropdowns
  • Attempted to support XML tags which are split during streaming

User Profiles

  • A default profile is provided and targets Ollama endpoint and no API key
  • Users can save new profiles and upon switching the API key and endpoint will be filled into the settings modal
  • Profiles are saved to disk in the userdata directory in a file named profiles.json

Additional Minor Improvements

  • Start scripts now check and warn of Node.js installation
  • Settings modal attempts to remain useable at high zoom levels
  • Improved API error handling hopefully prevents common backend crashes

Known Issues

  • As I don't currently have API for Anthropic models or OpenAI models their support is currently mostly theoretical so YMMV
  • Front and back end logging may be excessive and log levels aren't considered in any consistent way
Assets 3
Loading