Typing Mind

Can we please get custom system prompts for each model? Currently, it's only possible to set some settings unique for each model, but the system prompt is only global. It's frustrating if you want to maximize model potential.

Sorting chats resets back to "Recent Chats First" every time Typing Mind is refreshed. My request is that the sorting setting should be permanent (i.e. if I sort by Title A to Z, that sort is saved and permanent until I change it back to another sort), and not reset back to "Recent Chats First". As an alternative, add a keyboard shortcut option to match a sorting option to be able to apply a sort through keyboard shortcut.

Thank you for adding support for OpenRouter.ai models. It's really convenient. Are you planning to add similar functionality for image generation services like Replicate.com ? Currently, only DALL-E and Stable Diffusion are available, but that's not enough anymore. It would be great to be able to choose different image generation models ourselves, as new interesting models are constantly emerging.

so we dont have to keep updating version and be behind passively

Requesting this feature, cache pings like Aider does, Aider saves me a TON of money with this method, please implement it. "Preventing cache expiration Aider can ping the provider to keep your prompt cache warm and prevent it from expiring. By default, Anthropic keeps your cache for 5 minutes. Use --cache-keepalive-pings N to tell aider to ping every 5 minutes to keep the cache warm. Aider will ping up to N times over a period of N*5 minutes after each message you send." https://aider.chat/docs/usage/caching.html

The Gemini API has a new feature where you can ground results with Google search which reduces hallucinations and provides references to content. https://developers.googleblog.com/en/gemini-api-and-ai-studio-now-offer-grounding-with-google-search/ Here's the important description of the feature "the model returns grounding sources (in-line supporting links) and Search Suggestions that point users to the search results corresponding to the grounded response."

Currently there is an option to export the chat as MD, but would be good to export also only the message, not the full chat

TypingMind has made it very convenient to switch models part way through a chat session. However, it's exceedingly difficult to identify where such model switches took place. I think the most useful solution to this is to use the LLM's icon in the chat for the AI output, rather than the currect practice of using the TypingMind icon. This way, it is always clear and easy to identify which LLM produced any given AI output in the chats.

Create a TypingMind add-in for MS Word. The add-in would allow direct access to the AI models in Word. The chat should be saved in a folder and we should be able to open saved chats in the add-in. The add-in should have most of the capabilities as the current web interface—i.e., choose model, assign agent, use plugins, show chat info in upper right corner, attach documents, prompt library, send audio messages, etc. This would be an especially great feature for TypingMind custom. It provides a familiar environment for users to access and use the AI models. I have attached a mockup of what it might look like. Edited to note: based on my research, it appears that the MS Office add-in platform is powered by Edge webview2 (at least in Windows) so it should support the required web standards to implement this.