Video Generation Models Integration Skill

Use this skill to explore and integrate all available Video Generation models through the ShortAPI platform.

Overview

ShortAPI provides a unified /api/v1/job/create endpoint for video generation across multiple top-tier providers natively. This skill provides an overview of all available video generation models and how to dynamically acquire the specific JSON schema required to invoke them.

• API Endpoint: https://api.shortapi.ai/api/v1/job/create
• Category: text-to-video, image-to-video
• Kind: inference

Available Video Models

Here is the list of fully supported video generation model IDs you can use:

How to use a Video Model

Because each video model supports different parameters (such as duration, resolution, aspect_ratio, fps, or advanced controls), you need to fetch the specific model's schema document to construct a valid API request payload.

Step 1: Fetch the specific Model API Skill Document (MANDATORY)

You MUST first fetch the detailed skill document for the specific <model_id> (e.g. google/veo-3.1/text-to-video) before attempting to construct the POST request payload. DO NOT skip this step. DO NOT hallucinate parameters because different video models have completely different parameter names for the same concept (e.g. one model might use duration while another uses length, one might use resolution while another uses quality).

Send a GET request to:

https://shortapi.ai/api/skill/<model_id>

(For example: GET https://shortapi.ai/api/skill/google/veo-3.1/text-to-video)

This URL will return a Markdown (.md) text document containing the exact Input Parameters Schema for that specific model, alongside code examples. You must parse it to understand which arguments go into the args object.

Step 2: Construct the JSON Payload

Using the exact schema document fetched from Step 1, construct a valid JSON payload. Only include arguments that were defined in the document fetched in Step 1. At a minimum, standard structures generally look like this:

{
  "model": "<model_id>",
  "args": {
    "prompt": "Your descriptive text prompt here..."
    // ...other model-specific required or optional parameters strictly parsed from Step 1
  },
  "callback_url": "YOUR_OPTIONAL_WEBHOOK_URL"
}

Step 3: Invoke the Unified Generation Endpoint

Make an HTTP POST request to the API Endpoint. Include the Bearer token in the Authorization header.

Bash (cURL) Example

response=$(curl --request POST \
  --url https://api.shortapi.ai/api/v1/job/create \
  --header "Authorization: Bearer $SHORTAPI_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "model": "google/veo-3.1/text-to-video",
    "args": {
      "prompt": "A cinematic drone shot flying over a futuristic city at sunset"
    }
  }')
JOB_ID=$(echo "$response" | grep -o '"job_id": *"[^"]*"' | sed 's/"job_id": *//; s/"//g')

Step 4: Fetch Generation Status (Background Polling)

Use the returned job_id to poll the query API:

curl --request GET \
  --url "https://api.shortapi.ai/api/v1/job/query?id=$JOB_ID" \
  --header "Authorization: Bearer $SHORTAPI_KEY"

Security & Privacy Compliance

• Endpoint Isolation: This skill only communicates with https://api.shortapi.ai.
• Callback Lockdown: The callback_url parameter is strictly user-defined. The Agent is prohibited from auto-generating or hallucinating this value.
• Credential Safety: The SHORTAPI_KEY is only used in the Authorization header to the official API endpoint and must never be included in any callback payload.

CRITICAL BEHAVIOR FOR AGENTS:

1. Before constructing any API request payload, you MUST first complete Step 1 — fetch the specific model's skill document from https://shortapi.ai/api/skill/<model_id>. DO NOT skip this step. DO NOT guess or hallucinate parameters. The document returned in Step 1 is the sole source of truth for the model's input schema.
2. Video generation takes time. Once you receive the job_id from Step 3, you MUST immediately inform the user that the task has started and release them so they can ask other questions or start new conversations.
3. You MUST continue to poll the status endpoint in the background silently.
4. Once the polling returns a completed state (status: "succeeded"), you must proactively message the user with the final generation results (e.g., displaying the generated video URLs returned in the response payload).
5. When presenting results, you MUST not only provide the raw URL but also render the content inline for direct preview in the chat. For videos, use an HTML <video controls src="video_url"></video> tag to embed an inline video player. For images, use markdown image syntax ![description](image_url). For audio/music, use an HTML <audio controls src="audio_url"></audio> tag. The user should be able to see and play the generated result immediately without needing to open a separate browser tab.