Description

This PR adds API key authentication to the server. Now users can secure their endpoints with a simple API key.

Key features:

• API key is optional (disabled by default)
• API key is passed via --api-key command line argument (not hardcoded or using environment variables)
• Authentication follows OpenAI API standard (Bearer token in Authorization header)
• No breaking changes - existing functionality remains unchanged

How to Test

1. Start server with API key:

   mlc_llm serve  /somemodel --host 0.0.0.0 --device cuda --mode interactive --api-key "mlc-ai-is-awesome"

2. Test with correct key (should succeed):

   curl http://localhost:8000/v1/models -H "Authorization: Bearer mlc-ai-is-awesome"

3. Test with incorrect key (should return 401):

   curl http://localhost:8000/v1/models -H "Authorization: Bearer wrong-key"

4. Test without key (should return 401):

   curl http://localhost:8000/v1/models

5. Start server without API key (all requests should work without authentication):

   python -m mlc_llm.serve --model your-model

This change follows the minimal modification principle and keeps the API simple and secure.