clients:
  - type: openai-compatible
    name: ollama
    api_base: https://ollama.giugl.io/v1
    models:
      - name: pino 
        max_input_tokens: 8192
        max_output_tokens: 16000

      - name: pino-coder          
        max_input_tokens: 16000
        max_output_tokens: 16000

      - name: pino-embed
        type: embedding
        default_chunk_size: 512                        
        max_batch_size: 100

rag_embedding_model: ollama:pino-embed
rag_top_k: 5