feat(ai): add bge-large-en-v1.5 embedding model to litellm

- Add BGE embedding model config (port 8002) to litellm-config.yaml
- Add GPU_VLLM_EMBED_URL env var to compose and .env

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

ai/compose.yaml

@@ -104,6 +104,7 @@ services:
       GPU_TAILSCALE_IP: ${GPU_TAILSCALE_IP}
       GPU_VLLM_QWEN_URL: ${GPU_VLLM_QWEN_URL}
       GPU_VLLM_LLAMA_URL: ${GPU_VLLM_LLAMA_URL}
+      GPU_VLLM_EMBED_URL: ${GPU_VLLM_EMBED_URL}
       # LITELLM_DROP_PARAMS: 'true'  # DISABLED: Was breaking streaming
       NO_DOCS: 'true'
       NO_REDOC: 'true'

ai/litellm-config.yaml

@@ -55,6 +55,15 @@ model_list:
       supports_system_messages: true  # Llama supports system messages
       stream: true  # Enable streaming by default
 
+  # Embeddings - BGE Large (Port 8002)
+  - model_name: bge-large-en-v1.5
+    litellm_params:
+      model: hosted_vllm/BAAI/bge-large-en-v1.5
+      api_base: os.environ/GPU_VLLM_EMBED_URL  # Direct to vLLM embedding server
+      api_key: "EMPTY"
+      rpm: 1000
+      tpm: 500000
+
 litellm_settings:
   drop_params: false  # DISABLED: Was breaking streaming
   set_verbose: true  # Enable verbose logging for debugging streaming issues