* fix: Retrieve Gemini models by @yz778
* feat: Add Baidu search engine
* fix: Remove reasoning tag from TTS utterance
* fix: 403 issue when using Internet search.

- feat: added support for 'thinking collapsible' in reasoning models message
- fix: add error handling to fetch advanced Ollama settings data
- feat: change default search provider to DuckDuckGo
- fix: add query invalidation for sidebar settings after saving changes
- feat: no need for an embedding model for simple internet search

- feat: Added an option to download chats as images
- feat: Added an option to enable chat with a website by default
- feat: Added DeepSeek as provider
- fix: Resolved the model name issue when copying Markdown.

- feat: Add chat history and functionality to the side panel chat/copilot
- feat: Add option to enable internet search by default
- feat: Improve UI and UX

- feat: Added support for Mistral as a provider.
- feat: Added text splitting configuration options (`CharacterTextSplitter`, `RecursiveCharacterTextSplitter`).
- feat: Added additional Ollama model settings for `tfsZ`, `numKeep`, `numThread`, and `useMlock`.
- fix: Resolved various minor bugs.

- feat: Added OCR support. (Note: Currently, OCR supports English only. Additional languages and configurations will be introduced in future updates.)
- feat(model settings): Introduced support for minP, repeatLastN, and repeatPenalty.
- feat(tts): Added ElevenLabs as a TTS provider.
- feat(i18n): Added Arabic language support
- fix: Resolved various bugs and improved styling.

- feat: Change default side panel keyboard shortcut to Ctrl+Shift+Y #139
- feat: Add support for Google AI (Gemini) as a custom model provider
- feat: Add support for custom Google domains in search settings #71
- feat: Update Web UI

- Added a new system reset button.
- Integrated Brave Search API support.
- Improved the knowledge base to no longer track the original file after completion.
- Introduced new chat settings for copying (TXT, Markdown) and downloading (TXT, Markdown, JSON) the current chat.
- Fix: Resolved an issue where the system prompt was not saved in the first message.
- Fix: Switched from synchronous to local storage to avoid storage issues for certain settings.

- Bug fixes
- Added support for saving and restoring the last-used system prompt
- Fixed (Sidepanel): Enabled or disabled the Ollama connection status check.

- fix: Corrected spelling errors for Norwegian by Abubakar115e
- fix: Updated Ollama URL from `ollamaInfo` in `PlaygroundEmpty` by farooqpk
- feat: Added an option to resume the last chat when opening the Web UI
- feat: Introduced the `useMMap` option in model settings
- feat: Integrated the SearXNG search provider
- feat: Enabled auto-save functionality for the temporary system prompt in the current model settings drawer

- New chat model selection.
- Improved chat with website mode: Issues #228 and #234 addressed.
- Feature (Settings): Added Ollama connection status check setting.
- Feature (New Chat): Improved button styles and layout for RTL support.
- Ukrainian translation: Added Ukrainian translation for Page Assist by vlisivka
- [Experimental] Sidebar vision chat.

- Added Korean localization
- Added final generation info for Ollama
- Added max tokens setting for models
- Enabled sending images without text
- Updated chat model settings to use the selected system prompt if the temporary system prompt is empty
- Introduced temporary chat mode
- Added support for Llamafile
- Introduced multi-Ollama API support (beta)

* Swedish translations by @PierreMesure
* Fixed "Chat with PDF Content" issue in Firefox
* Fixed download button placement on the Hugging Face website
* Added: "Save" button for user message edits

- Updated Playground UI
- Added code download feature to code blocks
- Added pin/unpin option for chat history
- Grouped chat history by "Today," "Yesterday," "Last 7 days," and "Older"
- Added gguf model download button to Hugging Face website for the Ollama popup
- I18n fix for firefox

- Updated Playground UI
- Added code download feature to code blocks
- Added pin/unpin option for chat history
- Grouped chat history by "Today," "Yesterday," "Last 7 days," and "Older"
- Added gguf model download button to Hugging Face website for the Ollama popup

- OpenAI-compatible API support (easily connect LM Studio and Llamafile to Page Assist).
- Support for Danish and Norwegian languages.

* Added a "Clear All Chat History" button to the chat history drawer.
* [NEW] Added a temporary system prompt to the "Current Chat Model Settings" drawer.
* Fixed the page-share modal to use the correct title when sharing a chat.
* Bug fixes.

* Added LaTeX support to the markdown renderer
* Extended content script scope to cover all of ollama.com
* Added system and quick prompts to the side panel

* Added Deutsch
* Fixed IME issue of FireFox
* Added headers to every API call
* stt-to-storage: UX - Move speechToText from state to storage
* UI updated
* Bug fixes

* [FIX] Firefox sidebar style issues
* Added i18n to context menu items
* Added `num_gpu` parameter to Ollama model settings
* Added the ability to change the number of retrieved documents in RAG

- Added Persian (Farsi) language
- Added preset options to the context menu
- Added AI-generated title support
- Firefox Android Support

- Ability to turn off embedding for Chat with Website mode on the side panel.
- Ability to update the document limit for the knowledge base.
- Fixed issue of chat with website displaying irrelevant documents.
- UI fixes.

* added missing spanish translation
* fixed pt-BR translation issues
* set the current model to the last one used when opening a previous chat.
* Fixed auto-scroll issues on the playground.
* Send a notification after finishing the knowledge base (experiment)
* Added prevention of closing the page while processing the knowledge base
* UI updates