execute_template

Instructions

Implementation Reference

• src/yaparai/tools/templates.py:67-119 (handler)

  The main tool handler: async function execute_template that builds a payload, calls the client, and optionally polls for results.

  async def execute_template(
      slug: str,
      prompt: str,
      image_url: str | None = None,
      width: int = 512,
      height: int = 512,
      extra_inputs: dict | None = None,
  ) -> dict:
      """
      Execute a ComfyUI template to generate content.
  
      Runs a pre-built AI workflow with your inputs. Each template has
      different capabilities — use get_template_detail() first to see
      what inputs are accepted. Credits are deducted based on the template.
  
      Args:
          slug: Template slug (e.g., "flux-logo-generator", "product-photo-enhancer")
          prompt: Main text prompt for the template
          image_url: Input image URL (required for image-based templates)
          width: Output width in pixels (64-2048)
          height: Output height in pixels (64-2048)
          extra_inputs: Additional template-specific inputs (see get_template_detail).
                        e.g., {"brand_color": "#FF0000", "logo_text": "ACME Corp"}
  
      Returns:
          Dict with job_id, status, and result info.
      """
      client = YaparAIClient()
      payload: dict = {
          "prompt": prompt,
          "width": width,
          "height": height,
      }
      if image_url:
          payload["image_url"] = image_url
      if extra_inputs:
          payload.update(extra_inputs)
  
      result = await client.execute_template(slug, payload)
  
      # If the template returns a job_id, poll for result
      job_id = result.get("job_id")
      if job_id:
          final = await client.wait_for_result(job_id, timeout=120)
          return {
              "status": "success",
              "result_url": final.get("result_url"),
              "job_id": job_id,
              "credits_used": result.get("credits_used"),
              "balance_remaining": result.get("balance_remaining"),
          }
  
      return result

• src/yaparai/tools/templates.py:74-93 (schema)

  The docstring serves as the tool's schema/description, defining inputs (slug, prompt, image_url, width, height, extra_inputs) and return type.

  ) -> dict:
      """
      Execute a ComfyUI template to generate content.
  
      Runs a pre-built AI workflow with your inputs. Each template has
      different capabilities — use get_template_detail() first to see
      what inputs are accepted. Credits are deducted based on the template.
  
      Args:
          slug: Template slug (e.g., "flux-logo-generator", "product-photo-enhancer")
          prompt: Main text prompt for the template
          image_url: Input image URL (required for image-based templates)
          width: Output width in pixels (64-2048)
          height: Output height in pixels (64-2048)
          extra_inputs: Additional template-specific inputs (see get_template_detail).
                        e.g., {"brand_color": "#FF0000", "logo_text": "ACME Corp"}
  
      Returns:
          Dict with job_id, status, and result info.
      """

• src/yaparai/server.py:143-143 (registration)

  Registration of the execute_template function as an MCP tool via mcp.tool(execute_template).

  mcp.tool(execute_template)

• src/yaparai/server.py:51-52 (registration)

  Import of execute_template from yaparai.tools.templates into the server module.

      execute_template,
  )

• src/yaparai/client.py:175-177 (helper)

  The HTTP client method that sends the POST request to /v1/comfy-templates/{slug}/execute with the payload.

  async def execute_template(self, slug: str, payload: dict) -> dict:
      """Execute a ComfyUI template."""
      return await self._request("POST", f"/v1/comfy-templates/{slug}/execute", json=payload)