MCP DeepInfra AI Tools Server

Instructions

Implementation Reference

• src/mcp_deepinfra/server.py:217-239 (handler)

  The handler function for the 'text_classification' tool, registered via @app.tool(). It crafts a prompt for text classification (sentiment and category) using the configured model via DeepInfra's OpenAI-compatible completions API and returns the model's response.

      @app.tool()
      async def text_classification(text: str) -> str:
          """Classify text using DeepInfra OpenAI-compatible API."""
          model = DEFAULT_MODELS["text_classification"]
          prompt = f"""Analyze the following text and classify it. Determine the sentiment (positive, negative, neutral) and main category/topic. Provide your analysis in JSON format with 'sentiment' and 'category' fields.
  
  Text: {text}
  
  Response format: {{"sentiment": "positive/negative/neutral", "category": "topic"}}"""
          try:
              response = await client.completions.create(
                  model=model,
                  prompt=prompt,
                  max_tokens=200,
                  temperature=0.1,
              )
              if response.choices:
                  return response.choices[0].text
              else:
                  return "Unable to classify text"
          except Exception as e:
              return f"Error classifying text: {type(e).__name__}: {str(e)}"

• src/mcp_deepinfra/server.py:31-42 (helper)

  Configuration dictionary for default models used by tools, including the model for 'text_classification' (defaults to 'microsoft/DialoGPT-medium').

  DEFAULT_MODELS = {
      "generate_image": os.getenv("MODEL_GENERATE_IMAGE", "Bria/Bria-3.2"),
      "text_generation": os.getenv("MODEL_TEXT_GENERATION", "meta-llama/Llama-2-7b-chat-hf"),
      "embeddings": os.getenv("MODEL_EMBEDDINGS", "sentence-transformers/all-MiniLM-L6-v2"),
      "speech_recognition": os.getenv("MODEL_SPEECH_RECOGNITION", "openai/whisper-large-v3"),
      "zero_shot_image_classification": os.getenv("MODEL_ZERO_SHOT_IMAGE_CLASSIFICATION", "openai/gpt-4o-mini"),
      "object_detection": os.getenv("MODEL_OBJECT_DETECTION", "openai/gpt-4o-mini"),
      "image_classification": os.getenv("MODEL_IMAGE_CLASSIFICATION", "openai/gpt-4o-mini"),
      "text_classification": os.getenv("MODEL_TEXT_CLASSIFICATION", "microsoft/DialoGPT-medium"),
      "token_classification": os.getenv("MODEL_TOKEN_CLASSIFICATION", "microsoft/DialoGPT-medium"),
      "fill_mask": os.getenv("MODEL_FILL_MASK", "microsoft/DialoGPT-medium"),
  }

• src/mcp_deepinfra/server.py:216-217 (registration)

  Conditional registration of the text_classification tool based on ENABLED_TOOLS environment variable.

  if "all" in ENABLED_TOOLS or "text_classification" in ENABLED_TOOLS:
      @app.tool()