generate_video

Instructions

Implementation Reference

• src/yaparai/tools/generate.py:61-107 (handler)

  The main handler function for the generate_video tool. It accepts a prompt (text-to-video), optional image_url (image-to-video), a model selector (auto/veo/kling), and an optional style. Depending on the model choice and presence of image_url, it sets the 'mode' to 'gemini_video', 'img2video', or 'text2video', then calls the YaparAIClient to submit a generation job and polls for the result.

  async def generate_video(
      prompt: str,
      image_url: str | None = None,
      model: Literal["auto", "veo", "kling"] = "auto",
      style: Literal["cinematic", "realistic", "artistic"] | None = None,
  ) -> dict:
      """
      Generate a video using AI.
  
      Text-to-video or image-to-video. Models: Veo 3.1, Kling.
      Cost: ~350 credits per video.
  
      Args:
          prompt: Description of the video to generate
          image_url: Optional source image URL (for image-to-video mode)
          model: Model to use — "auto" (default, picks best), "veo" (Veo 3.1),
              or "kling" (Kling). Veo 3.1 is best for cinematic quality.
          style: Style preset (cinematic, realistic, artistic)
  
      Returns:
          Dict with video_url, job_id, credits_used, and balance_remaining.
      """
      client = YaparAIClient()
  
      if model == "veo":
          mode = "gemini_video"
      elif image_url:
          mode = "img2video"
      else:
          mode = "text2video"
  
      job = await client.generate({
          "type": "video",
          "prompt": prompt,
          "mode": mode,
          "image_url": image_url,
          "style": style,
      })
  
      result = await client.wait_for_result(job["job_id"], timeout=180)
      return {
          "status": "success",
          "video_url": result.get("result_url"),
          "job_id": result.get("job_id"),
          "credits_used": job.get("credits_used"),
          "balance_remaining": job.get("balance_remaining"),
      }

• src/yaparai/tools/generate.py:61-66 (schema)

  The function signature and docstring for generate_video, defining input parameters (prompt, image_url, model with Literal types, style) and the return type dict with video_url, job_id, credits_used, balance_remaining.

  async def generate_video(
      prompt: str,
      image_url: str | None = None,
      model: Literal["auto", "veo", "kling"] = "auto",
      style: Literal["cinematic", "realistic", "artistic"] | None = None,
  ) -> dict:

• src/yaparai/server.py:26-26 (registration)

  Import of generate_video from yaparai.tools.generate into the server module.

  generate_video,

• src/yaparai/server.py:124-124 (registration)

  Registration of generate_video as an MCP tool on the FastMCP server via mcp.tool(generate_video).

  mcp.tool(generate_video)

• src/yaparai/client.py:126-128 (helper)

  The YaparAIClient.generate method used by generate_video to POST to /v1/public/generate with the job payload.

  async def generate(self, request: dict) -> dict:
      """Start a generation job."""
      return await self._request("POST", "/v1/public/generate", json=request)