Add integrated LLM chat mode to Walker

[Rishabh-Sarang]

Describe the feature
Add an optional built-in LLM chat mode inside Walker.
Users could type natural-language questions (e.g., “What’s the weather?”, “Explain recursion”, “Summarize this text”) directly into the launcher and get answers from an LLM.

This feature would make Walker a lightweight, fast-access AI assistant — no need to open a separate browser or app.

The LLM backend should be configurable, supporting:

• Remote APIs (e.g., OpenAI, Anthropic, etc.)

Describe the behaviour

• Activation:

  1. User opens Walker as usual.
  2. Presses a hotkey or enters a trigger (e.g., /ai or ?) to enter Chat Mode.

• Input:

  1. User types a natural-language prompt, e.g.:
  2. “/ai What is the fastest sorting algorithm?”
  3. “Explain the difference between TCP and UDP.”

• Processing:

  1. Walker sends the prompt to the configured LLM backend.
  2. The backend could be local (Ollama, LM Studio) or API-based (OpenAI, etc.), depending on user settings.

• Output:

  1. The model’s response is displayed directly in Walker — either inline or in a small popup/chat pane.
  2. User can press Shift+Enter (or similar) to continue the conversation contextually.
  3. Optionally allow copying the response with one key.