Skip to content
RecurseChat

FAQ

1.4.11

  • Fixed an issue with Llava when upgrading from an older version
  • Updated help menu, include discord link

1.4.9

  • Added feature to organize model in folders
  • Performance improvements
  • Visual polish and updates

1.3.8

  • Fixed an issue with Azure OpenAI integration.

1.3.7

  • Fixed an issue with fade-in streaming animation in lists.

1.3.4

  • Light mode Code Color Theme now defaults to a light theme.
  • Added Code Color Theme setting.

1.3.2

  • Added Floating Chat feature. You can now chat in a floating chat window. Activate the floating chat window with keyboard shortcut ⌘⇧↩ or from the status bar (menu bar). You can customize the keyboard shortcut in Settings.

1.2.2

  • Performance improvements

1.1.12

  • Added support for Llama 3.2 1B and Llama 3.2 3B

1.1.11

  • Fixed an issue with LaTeX rendering in PDF exports

1.1.10

  • Added support for LaTeX rendering

1.1.8

  • Added option to regenerate response with a different model

1.1.5

  • Fixed an issue with exporting conversation to PDF

1.1.4

  • Bookmarks: Right click on a chat session to bookmark it from the options button next to the leading sidebar
  • Folders: You can create folders, expand all / collapse all folders, choose different item sizes from the dropdown button located next to the leading sidebar button
  • Multiple selection: You can select multiple items to reorder drag, bookmark or delete
  • Performance improvements

1.0.156

  • Added Llama 3.1 model
  • Added Gemma 2 model
  • Upgraded llama.cpp to b3502
  • Added configuration for context size used in llama.cpp
  • Added Claude models

1.0.153

  • Upgraded llama.cpp version
  • Added model’s GGUF metadata information in advanced settings
  • Use model’s context size from metadata
  • Added support for GPT-4o mini

1.0.149

  • Updated message bubble styling. You can customize message UI appearance by clicking the dropdown on the toolbar.
  • Added font choices
  • Style improvements

1.0.141

  • Added feature to edit and branch messages. Newly imported ChatGPT history should also have branches preserved.
  • Fixed an issue with deleting a duplicated model

1.0.134

  • Add a files setting for models. You can now specify multiple files for a model.
  • Improved citation hover view
  • Added fade-in animation for streaming. You can configure this setting in Settings -> Appearance.

1.0.126

  • Added option to restore all default models. If you would like to restore the built-in default models, go to Settings -> Models -> Restore default models and click “Restore all”.

1.0.124

  • Added feature to duplicate models: Right click a model in the model page to duplicate the model
  • Added GPT-4o as a built-in model

1.0.120

  • Fixed an issue with indexing large PDF files

1.0.119

  • Supported adding model directly from Hugging Face. Go to Model Page -> New Hugging Face GGUF Model to create a model directly from a Hugging Face URL.
  • Improved chat templates page and added default presets.

1.0.116

  • Fixed an issue with indexing certain long pdfs

1.0.115

  • Fixed an issue with qdrant server shutdown

1.0.113

  • Fixed an issue where user is able to drag & drop image in chats with text-only models

1.0.112

  • Improved chat input styling
  • Improved search input styling

1.0.109

Upgrade llamacpp version

1.0.110

  • Upgraded llama.cpp version
  • Better handling of small window size
  • Polished settings UI, added option to delete all chats

1.0.99

  • Updated GPT-3.5-turbo and GPT-4-turbo model
  • Added keyboard shortcut for deleting chat
  • UI polishes and better handling of initial embedding model downloading.

1.0.93

  • improves dark mode aesthetics
  • add light / dark mode theme selection in addition to system theme
  • add embedding model home page and license in the info bubble
  • adjusted order of model categories’