Add a model selector at menu

[ysaereve]

Is your feature request related to a problem? Please describe it clearly. N/A

Describe the feature you'd like to see implemented. Add a model selector to the menu, e.g. allowing users to choose a model with or without RAG support. In the settings, provide options to configure the primary, secondary, and tertiary models for the connection. The model selector should have the same background color as the connection it's associated with.

Why is this feature valuable? Let users choose the model that works best for the prompt (workload).

Additional context, mockups, or screenshots

[micz]

I've added the new feature label. This won't be easy.

[ysaereve]

Thanks so much for considering this! I’d love to help with testing when you have a prototype, so please feel free to reach out.

[micz]

Given that I'll implement #102, so you could define the API and model for each prompt, do you think the selector in the menu will still be needed?

[ysaereve]

Thank you for giving me the opportunity to share my feedback. I hope the following is helpful.

API Provider + Model Selector

Task Level vs. Prompt Level

---

Task Level

Pros:

• Gives users the freedom to choose their preferred API Provider + Model for a specific task.
• Makes it easy to switch between local and online provider + model options.

Cons:

• Users must remember which provider + model is best suited for their specific prompt.

---

Prompt Level

Pros:

• Prompts can be tailored to work optimally with specific provider + model combinations.
• Enhances performance for tasks that are highly dependent on a particular provider + model.

Cons:

• Prompts are static and not dynamic, limiting flexibility.
• Requires separate prompts for different provider + model combinations.
• May necessitate grouping prompts for efficient management.

---

Comparison Table

Feature	Task Level	Prompt Level
Flexibility	High	Low
User Control	High	Low
Performance	May vary	Optimized for specific provider + model
Management Complexity	Low	High

---

Conclusion

By default, implementing the Prompt Level feature, while retaining Task Level as a future option to override the Provider + Model selector. This approach allows users to benefit from optimized performance for their specific tasks, while still maintaining flexibility and control over their provider + model choices.

---

[micz]

Thank you for your feedback. I'll keep this issue open for future consideration as a new feature.

[ysaereve]

That's great! Thank you very much!