2.4.37

• The Query only mode in Uploaded tab has been renamed to RAG.
• New options have been added under Settings -> Files and Attachments:
  • Use history in RAG query: When enabled, the content of the entire conversation will be used when preparing a query if the mode is set to RAG or Summary.
  • RAG limit: This option is applicable only if 'Use history in RAG query' is enabled. It specifies the limit on how many recent entries in the conversation will be used when generating a query for RAG. A value of 0 indicates no limit.
• Cache: dynamic parts of the system prompt (from plugins) have been moved to the very end of the prompt stack to enable the use of prompt cache mechanisms in OpenAI.

2.4.36

• Added a new command-line argument: --workdir="/path/to/workdir" to explicitly set the current working directory.
• Fix: start image generation in Image mode.

2.4.35

• Docker removed from dependencies in Snap version #82
• Refactored documentation.
• Fix: toggles real-time update hook.
• Fix: missing edit icons.
• Added tokens from attachments to counters if mode == Full context.

2.4.34

• Added a new mode: Chat with Audio, with built-in multimodal support for audio input/output. Currently in beta, the execution of commands and tools in this mode is temporarily unavailable.
• Added new models: gpt-4o-audio-preview, gpt-4o-2024-11-20, chatgpt-4o-latest.
• Force disabled integration with the native system menu.

2.4.33

• Improved CSS and rendering of file and image lists.
• Added displaying of used attachments in the chat window.

2.4.32

• The "Add URL" option added to the "Attachments" tab allows users to include content from a given website as additional context. Currently, it only supports standard web pages and video transcription for YouTube links. More "web" options will be added in the future.
• Added UTF-8 as default in attachments content text read/write.

2.4.31

• Added an option checkbox Auto-index on upload in the Attachments tab:

Tip: To use the Query only mode, the file must be indexed in the vector database. This occurs automatically at the time of upload if the Auto-index on upload option in the Attachments tab is enabled. When uploading large files, such indexing might take a while - therefore, if you are using the Full context option, which does not use the index, you can disable the Auto-index option to speed up the upload of the attachment. In this case, it will only be indexed when the Query only option is called for the first time, and until then, attachment will be available in the form of Full context and Summary.

• Added context menu options in Uploaded attachments tab: Open, Open Source directory and Open Storage directory.

2.4.30

• Added instruction to model about mapped data directory in both legacy and IPython code interepreter.
• Updated locales for plugins.

2.4.29

• Added a local IPython interpreter - you can now choose between local and sandbox (Docker) in the plugin settings.
• Added the ability to configure mapped volumes and ports for Docker containers in the plugin settings.
• Optimized and speed-up the Llama-index plugin (inline).
• Checkboxes replaced with Toggle buttons.
• Improved settings dialogs.
• Slight modification of the layout.
• Fix: Dockerfile formatting in Code Interpreter config.
• Fix: experts inline plugin execution.

2.4.28

• Fix: issue #78