Embedding coordinator

Lightweight embedding generation for Erlang. Manages a chain of embedding providers with automatic fallback.

Providers

Local: - local: Local Python with sentence-transformers (CPU, no external calls) - ollama: Local Ollama server - fastembed: FastEmbed ONNX-based embeddings (lighter than sentence-transformers)

Cloud: - openai: OpenAI Embeddings API - cohere: Cohere Embed API - voyage: Voyage AI Embeddings - jina: Jina AI Embeddings - mistral: Mistral AI Embeddings (EU data residency) - azure: Azure OpenAI Embeddings - bedrock: AWS Bedrock Embeddings (Titan, Cohere) - vertex: Google Vertex AI Embeddings

Specialized: - splade: SPLADE sparse embeddings for hybrid search - colbert: ColBERT multi-vector embeddings for fine-grained matching - clip: CLIP image/text cross-modal embeddings

Configuration

   %% Local Python embeddings (requires Python + sentence-transformers)
   Embedder = {local, #{
       python => "python3",
       model => "BAAI/bge-base-en-v1.5"
   }}.
  
   %% Ollama server
   Embedder = {ollama, #{
       url => <<"http://localhost:11434">>,
       model => <<"nomic-embed-text">>
   }}.
  
   %% OpenAI API (requires API key)
   Embedder = {openai, #{
       api_key => <<"sk-...">>,  %% or set OPENAI_API_KEY env var
       model => <<"text-embedding-3-small">>
   }}.
  
   %% Provider chain with fallback
   Embedder = [
       {ollama, #{url => <<"http://localhost:11434">>}},
       {local, #{}}  %% fallback to CPU
   ].