TypingMind (→ feedback.typingmind.com)

Have centralized tag management.

Current search function is covering the whole chat history. I already sorted my chats in specific folders. But then then search function will search my whole history of chats, that include other folders too. Can we make the search is within active folder?

The folder system is to simple. It would be helpful to get subfolders (folder within folders)

Hey! ## Concern: Currently, when in the chat interface, opening the prompt library causes any text in the message field to be erased. The mechanism used doesn't seem to allow for ctrl-z undo. ## Behavior: User writes full prompt User opens Prompt Library modal for a potential alternative → Prompt in progress is erased. User decides against using a Library prompt after assessing User's writing is erased with no means of recovery --- ## Idea: I understand that this is likely a preparation step for inserting the Library prompt, but is there a way to mitigate this? ### Potential Behavior 1: User writes full prompt User opens Prompt Library modal for a potential alternative User browses and selects prompt from Library and clicks Use → Prompt in progress is erased on Library prompt selection, Library prompt is inserted ### Potential Behavior 2: User writes full prompt User opens Prompt Library modal for a potential alternative User browses and selects prompt from Library and clicks Use If text field empty -> Insert prompt If text field has content -> Show dialog: [ Replace ] [ Append ] [ Cancel ] Understandable if not possible, just wanted to toss this out there. Typically I draft in a text app but I always appreciate when text fields have high confidence persistence. TM is great at this otherwise. Thanks so much for all your work.

It would be a simple but powerful addition to the chat interface if we could have a full text editor (italicize, underline, bold, bullet points, indentation, etc.) to better format and organize our prompts/input. The models regularly return well-formatted text, but we are restricted to simple blocks of text seperated by line breaks. This should be improved!

When I try to copy model responses into a .docx or .txt file, I can only seem to copy/paste the raw text formatting. This effectively breaks the copy/paste functionality any time you are trying to move text out of TypingMind, because the raw formatting is virtually unreadable for longer, more detailed responses. Just a few weeks ago, I was able to do this, but now it is broken for some reason. Please fix ASAP!

Currently, TypingMind allows users to set the streaming output ON or OFF when adding Gemini models as custom models. However, the official built-in Gemini models provided by TypingMind do not have a streaming toggle in their settings. I would like to request that the official built-in Gemini models (e.g., Gemini 1.5 Pro, Gemini 2.0 Flash) also support enabling or disabling streaming output individually, just like custom models. In addition, I would like to have an option to enable or disable the streaming output when creating or editing AI Agents. This would allow users to control the streaming setting for each agent’s responses, resulting in a more flexible and consistent user experience. This would make the user experience more consistent and allow more flexible control over the output display style for both models and AI agents.

It would be great to have a keyboard shortcut to focus keyboard controls onto the message window (where the list of all messages appear). I couldn't find such a shortcut in the TypingMind docs ( https://docs.typingmind.com/general-settings/keyboard-shortcuts ). Currently after sending a message and receiving a response, if you want to then scroll through the response (if it is longer than the viewport) then you need to take your hand off the keyboard and onto the mouse. This can become tedious after a while, or even difficult, depending on the condition of your wrist, hand, etc. We have the / shortcut that lets you focus back on the text-input field after having scrolled the messages pane, but we don't have one that lets you focus on the messages. Even if you want to scroll up-down with the keyboard arrow keys, you first still need to use the mouse to select the message pane. Suggested possible keys: Esc key (intuitive) backtick ` key tilde ~ key Cmd/Ctrl + / key (to alternate with the same key that focuses text-input). Or even better: up key to focus on the messages so you can simply press up to start scrolling. And then when you reach the bottom, have the down key focus on the text-field again, for a seamless experience. This would require changing the shortcut for "edit previous message" though. ## Alternate default setting? For users who regularly receive responses that are taller than the viewport, it would even be helpful if the default behaviour upon receiving a response was to focus on the message pane, since it will always be necessary to scroll it to read the full message. This would mean that a workflow would only ever require using the / keyboard to focus on the text-input after reading a response (or maybe not even that, if the user's next response is to switch app). Otherwise the workflow will require both keys: one to focus on the messages, and then another / to go back to the text-input.

This would allow us to search and interact with web content. The alternative is to create a Web Browser extention for the major browsers.

Finding a specific model in the huge list of "Import Open Router" models is extremely cumbersome. A filter (preferably with wildcards) for the "Name" column would be enormously helpful!