Skip to main content

Quick Switch Model Parameter Presets (“Chat Modes” configuration)

Feature Request

It would be extremely helpful to create and save different “modes” or “presets” for model parameters (temperature, max_tokens, top_p, context, caching, etc.), and quickly switch between them in any chat with a single click, instead of adjusting every setting manually each time.

Motivation/Benefits:

  • Allows users to define several work styles: e.g., Economic Mode (for quick queries), Balanced Mode (for standard tasks), Deep Mode (for critical analysis).

  • Saves time and reduces the risk of forgetting to change a parameter.

  • Makes it much easier to control costs and response depth by toggling preconfigured modes.

  • Especially useful for power users, teams, consultants, or anyone that combines workflows of different depth and cost.

Example of use:

  1. The user creates and saves various presets (“Economic”, “Deep”, “Creative”, etc.).

  2. In each chat, a quick menu allows switching between modes; all parameters update instantly.

  3. This way, work style and cost management become much easier and more reliable.

Thank you for considering this suggestion! This would really improve TypingMind for those of us who work intensively with different depth, creativity, and cost needs.

3 comments

Log in to comment and vote

Comments3

  • Ngoc Nguyen

    Team•

    Dec 8, 2025

    As of now, you can configure different sets of parameters through the agent settings and switch between multiple AI agents within the same conversation. It should work.

    • Yellow Lamp

      •

      Dec 19, 2025

      Thank you very much for the feedback and for clarifying the current workflow with the AI ​​agents and their parameter presets. I appreciate the flexibility this method offers and use it to switch between different "modes" by configuring different agents.

      My suggestion (for the future) would be to make it even more efficient: It would be very helpful to be able to adjust the model parameters on the fly during a conversation and, if we find a combination that works particularly well, instantly save the current set of parameters as a new preset or mode, without having to manually create or clone agents each time. This would simplify experimentation and allow advanced users to build a library of useful parameter sets directly from practical experience in real conversations.

      Thanks again for your work and for your openness to feedback.

  • Plum Grove

    •

    Dec 6, 2025

    So clone or copy with new name and tweak settings for each scenario?

    I was thinking this plus on the fly changing of model parameters mid chat. This would allow iterative changes and when you are happy with the results save that variant of model settings with a new name.