[feature] Why don't support remote embedding services

[yiakwy-xpu-ml-framework-team]

It is very silly to use local cpu to run embedding models.

Simple configure a remote gpu endpoint (we can support both v1/chat/completions nad v1/embeddings) will be very convient for us.

Where can I add the options ? @tobi

node@3f218953d795:/app$ qmd status
QMD Status

Index: /home/node/.cache/qmd/index.sqlite
Size:  3.7 MB

Documents
  Total:    68 files indexed
  Vectors:  118 embedded
  Updated:  15m ago

Collections
  local (qmd://local/)
    Pattern:  **/*.md
    Files:    48 (updated 15m ago)
  openclaw-engram (qmd://openclaw-engram/)
    Pattern:  **/*.md
    Files:    0 (updated never)

Examples
  # List files in a collection
  qmd ls local
  # Get a document
  qmd get qmd://local/path/to/file.md
  # Search within a collection
  qmd search "query" -c local

Models
  Embedding:   https://huggingface.co/ggml-org/embeddinggemma-300M-GGUF
  Reranking:   https://huggingface.co/ggml-org/Qwen3-Reranker-0.6B-Q8_0-GGUF
  Generation:  https://huggingface.co/tobil/qmd-query-expansion-1.7B-gguf
QMD Warning: no GPU acceleration, running on CPU (slow). Run 'qmd status' for details.

Device
  GPU:      none (running on CPU — models will be slow)
  Tip: Install CUDA, Vulkan, or Metal support for GPU acceleration.
  CPU:      4 math cores

Tips
  Add context to collections for better search results: local, openclaw-engram
    qmd context add qmd://<name>/ "What this collection contains"
    qmd context add qmd://<name>/meeting-notes "Weekly team meeting notes"
  Add update commands to keep collections fresh: local, openclaw-engram
    qmd collection update-cmd <name> 'git stash && git pull --rebase --ff-only && git stash pop'
node@3f218953d795:/app$ cat ~/.config/qmd/index.yml 
collections:
  local:
    path: /home/node/.openclaw/workspace/memory/local
    pattern: "**/*.md"
  openclaw-engram:
    path: /app/openclaw-engram
    pattern: "**/*.md"
embedding:
  provider: openai
  base_url: http://host.docker.internal:8005/v1
  model: bge-m3
  dimensions: 1024
  api_key: dummy
logging:
  level: info
node@3f218953d795:/app$