- Improved markdown formatting - Output text renderers moved to separated modules - Added plain-text render option in settings - Vision inline disabled in unsupported modes (Image generation and Assistant) and hidden in Vision (always enabled here) - Updated OpenAI Chat Completion API
2.0.77
- Fixed inline plugins disable when pausing command plugins by unchecking 'Execute commands' checkbox - Added real-time image appending into chat window when generating images in Image Generation mode - Added automatic creation of context after deletion from the list when no other context is selected - Updated docs
2.0.76
- Markdown post-process changed from markdown extension to BS4 html parser.
2.0.75
- Improved default theme for markdown styling - Added "Check for updates on start-up" config option in updater dialog - Added custom CSS stylesheets editor in Config menu - Added password show/hide in password/secret input fields (click on the right corner of input reveals plain text) - Fixed JSON files editing (added data reloading after save to prevent overwrite on exit) - Added tests for plugins and extended tests for core/controllers
2.0.74
- Fixed timestamp position when appending input to the chat window. - Extended the Markdown parser with an extension for converting `li` to `p` (it allows copying lists bullets with Ctrl-C). - Added notebook titles to the "Copy to..." context menu. - Added an "Open file" option in the file explorer and attachments list.
2.0.73
- Fixed vision prompt appending - Improved inline vision: removed the "keep" option and added a vision indicator with a checkbox for quickly enabling/disabling the vision model if needed (will auto-enable when an image is provided if the plugin is active)