/claim #1846 Fixes #1846

Adds DeepSeek provider support for LLM Proxy and Chat.

Implementation follows the same pattern as other OpenAI-compatible providers (vLLM, Cerebras, ZhipuAI).

What’s included

• LLM Proxy routes and adapter
• Chat integration with model listing
• Type definitions and error handling
• Frontend UI components

Testing

Verified with DeepSeek API - models list correctly, streaming and non-streaming completions work.

Demo