v1.10.0 RC2

What's new

• Added a unified search agent class to reuse the script for each agent.
• Added Gemini chat model and embedding model provider.

Fixes

• Unhandled rejection caused by empty fileIds field.

v1.10.0 RC

What's new

• Added file upload functionality in webSearch focus mode. Currently, if any other search mode is selected other than web search, it will automatically change it to web search (As implementing file uploads in other focus modes is left, it will be added by the meta search agent class which will be introduced in the next version)
• Added keep_alive variable to specify the time for how long the model should be loaded in Ollama (Defaults to '5m')

v1.9.3

What's new

• Automatic focus on opening the site (on mount).

Fixes

• markdown-to-jsx package has been updated to fix issues with unclosed code block causing client side errors.
• The sources for summarization now have a maximum limit of 15 to prevent context errors.
• Downgraded/fixed NodeJS Docker version to prevent build errors caused by mismatch in package files.

v1.9.2

What's new

• Settings menu has been implemented on mobile devices

Fixes

• Delete and new chat button in the navbar have been fixed
• window.location.href is now used for refreshing the page (A new chat with plus button wasn't refreshing since its the same route)
• Custom OpenAI provider has been fixed in video search, image search and suggestions generation.

v1.9.1

Fixes

• Fixed rewrite not working due to duplicate message IDs
• Fix infinite loading if no models configured (for custom OpenAI users)

v1.9.0

What's New

• Add support for using Perplexica via an API endpoint.
• Added URL & PDF searching capabilities.
• Added line output parser with handling for invalid keys.
• Handle unhandled exceptions and rejections for improved stability.
• Added prebuilt Docker images with multi-architecture support and caching to improve build times.
• Integrated support for Llama 3.2.
• Added optimization modes to search agents and search API.
• Enhanced the chat window to only send the initial message when ready.
• Introduced a new discover feature in the app.
• UI enhancements for the library section.
• Added displayName property to providers for better clarity.
• Linked config.toml as a volume in Docker Compose.

Fixes

• Fixed infinite loading by closing socket on unmount.
• Updated the WebSearchAgent to fix typo and optimize prompt handling.
• Fixed issues with unresolved types across various modules.
• Prevent WebSocket errors when not open.
• Fixed the inability to type / in the settings dialog.
• Added a check for empty arrays to avoid client errors on image and video displays.
• Fixed large SearXNG settings file issues.
• Updated build workflows to improve stability and consistency.
• Fixed q search param not working due to WebSocket issues.

v1.9.0 RC3

What's new

• Search API to use Perplexica via an API endpoint.
• Changed models in each provider to object; add new display name property.
• Link config.toml file as a volume
• Increase sources quality by using AI generated query (the variable existed but I returned the actual question instead of the AI generated question)

Fixes

• Fix infinite loading by closing socket on unmount.
• Prevent Web search agent from sending example links for summarization (as passed in the prompt) & cause moving disc errors; update prompt & change temperature.
• Fix not able to type / in settings dialog.
• Check for empty arrays on images & videos to prevent client errors

v1.9.0 RC2

Fixes

• Unresolved types at some places
• Infinite loading in the UI
• Large SearXNG settings file

v1.9.0 RC

What's new

• Add support for searching & summarizing links and PDFs (from links).
• Add line output parser

Fixes

• Fix unhandled exceptions & rejections by catching errors.

v1.8.1

Release notes

Fixes

• Fixes undefined model for custom OpenAI provider
• Fixes theme for toast container.

Additions

• Add support for Llama 3.1
• Optimize document filtering and sorting