Fixed compatibility issues with OpenRouter integration following their API update today. Adjustments were made to align with their new API structure.
- New UI. This is a major update. Please make sure to clear your entire cache.
- Fixed AI Forms new line issue.
- Improved memory management for the chatbot, allowing higher memory limits for longer conversation history.
- Fixed an issue where the bot occasionally failed to remember previous conversations.
- Now we are sending user queries via form submission instead of the URL, removing URL length limitations in the chat bot while using streaming.
- Fixed encoding issues for non-Latin languages like Ukrainian and Russian.
- Export logs now verify proper permissions for the uploads folder.
- Improved handling of the o1 model by bypassing unsupported parameters such as top_p, frequency and presence penalties, and streaming. Check o1 models limitations [here](https://docs.aipower.org/docs/ai-engine/openai/gpt-models#o1-models-beta).
[New]
- Added support for new image models: Flux, SDXL, Kandinsky, and more for text-to-image generation across Content Writer, AutoGPT, and Image Generator modules.
For more details, refer to the [documentation](https://docs.aipower.org/docs/content-writer/express-mode/images#replicate)
[Improvements]
- Better markdown support for chat bot.
- Fixed broken links in chat response.
**Note:** If you're using a caching plugin, please remember to clear your site cache to ensure that the changes take effect.
- Now you can revise chatbot responses that didn't meet your expectations. Use this feature to improve future interactions.
[Read tutorial](https://docs.aipower.org/docs/ChatGPT/advanced-setup/context#response-quality)
- Added detailed stats including log size, conversation count, question count, and insights into the most popular models and providers.
[NEW]
- Added Internet Browsing feature for the chat bot.
Learn how to configure and enable it in our [tutorial](https://docs.aipower.org/docs/ChatGPT/advanced-setup/internet-browsing).
- Added support for GPT-4o-mini. Please go to **Settings - AI Engine** tab and hit Sync button to get the latest models from OpenAI.