[Feature Request] Dynamic Model/Combo Picker via '/v1/models' (Settings Dropdown + Chat Footer Quick Switcher)

[alenguav]

Problem Statement:

Currently, using the OpenAI Compatible provider requires manually typing Model IDs or alias strings into a plain text input field in Settings. Switching between models or routing profiles requires navigating away from the chat, manually re-typing the ID, and saving.

When working with local gateways and proxies (OmniRoute, LiteLLM, OneAPI):

1. Users constantly alternate between fast/cheap models (for reading files and scanning context) and reasoning-heavy models (for complex debugging).
2. Gateways expose Custom Routing Combos/Aliases (e.g., cline-coder, gemini, mistral, claude) via the standard /v1/models endpoint, but users have to type them manually in settings every time.

---

Proposed Solution:

Integrate dynamic model/combo fetching via the /v1/models endpoint across two key UI touchpoints:

1. Native Settings Dropdown:
   Replace the static text input for OpenAI Compatible providers with a dynamic dropdown populated by querying GET {Base URL}/v1/models. Treat gateway Combos/Aliases as native, selectable models.
2. Chat Footer Quick Switcher:
   Add a quick Model/Combo dropdown directly in the chat panel footer (next to the chat input/mode switcher).
3. One-Click Switching:
   Allow users to switch the active model or router combo on the fly before sending the next prompt—without opening the global Settings view.
4. Persist per Session:
   Changing the model via the chat UI applies immediately to the active task context.

---

Gateway / Router Use Case (OmniRoute, LiteLLM):

When connected to an OpenAI-compatible gateway like OmniRoute:

• The /v1/models endpoint returns the user's defined combos: cline-coder (fallback chain), gemini (fast/cheap), mistral (code fixer), claude (deep reasoning).
• The Workflow: A developer starts a task with the gemini combo for cheap scanning. When hitting a stubborn bug, they switch to the mistral combo in 1 click directly from the chat box to resolve the issue, then switch back—without leaving the chat context or typing alias names.

---

Benefits:

• Drastic Cost Optimization: Encourages users to use cheaper models/combos for lightweight tasks and switch to heavier ones only when needed.
• Zero Typing Errors: Eliminates typos in long Model IDs or custom proxy alias names.
• Seamless UX: Keeps the developer focused in the chat flow without disrupting their context.