Seedance 2.5 API and MCP: how to add video generation to your project

Seedance 2.5 is available on NeuralSpace not only in the web interface, but also through the public /v1 API and the Model Context Protocol (MCP) server. Developers and AI agents can generate videos directly in code: submit text prompts, pass reference images and videos, choose durations, and fetch the finished clip via URL.

How the API works

Video takes minutes to generate, so the API is asynchronous. First you send a POST request to create a task — the service calculates the cost, charges tokens immediately and returns a task identifier. Then you poll the task status, and when the clip is ready you receive a direct link to the video file. On error or timeout, tokens are refunded automatically. API keys can be generated in your dashboard under API Keys.

API Endpoints

Create a task with POST /v1/videos/generations using { model: "seedance-2.5", prompt, duration, resolution, aspect_ratio, image_urls, video_urls, audio_urls }. Check status and fetch the result via GET /v1/videos/generations/:id. The full catalog of models and live rates is documented at API Documentation and via GET /v1/models.

Connecting via MCP (Model Context Protocol)

For Claude Desktop, Cursor, Cline, and other AI agents, NeuralSpace hosts an official MCP server at https://neuralspace.pro/mcp. Seedance 2.5 is supported in the generate_video tool, with progress tracked via check_video. An AI agent can formulate prompts, configure parameters, and attach reference files autonomously.

{
  "mcpServers": {
    "neuralspace": {
      "url": "https://neuralspace.pro/mcp",
      "headers": {
        "x-api-key": "nsk-your_key"
      }
    }
  }
}

Seedance 2.5 Capabilities

In both the public API and MCP, the model is available under the ID seedance-2.5. It supports 480p, 720p, and 1080p resolutions, durations from 4 to 30 seconds, audio generation, aspect ratios (16:9, 9:16, 1:1, 4:3, 3:4, 21:9, or adaptive), as well as reference images and videos. You can test all parameters before coding on the model page.

Sample cURL Request

Creating a generation task via the public API:

curl https://neuralspace.pro/v1/videos/generations \
  -H "x-api-key: nsk-your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedance-2.5",
    "prompt": "close-up of a futuristic crystal on a dark table, soft neon glow, cinematic motion",
    "duration": 5,
    "resolution": "720p"
  }'

The response returns a task id, pending status, and polling URL. To retrieve status:

curl https://neuralspace.pro/v1/videos/generations/TASK_ID \
  -H "x-api-key: nsk-your_key"

When completed, the response includes the video URL in data[0].url.

FAQ

Is Seedance 2.5 available through API and MCP?

Yes, Seedance 2.5 is fully available both in the public REST API (model ID seedance-2.5) and through the official MCP server via the generate_video tool. Full parameter documentation is available in the API Documentation.

How do I get an API key?

Keys are issued instantly in the NeuralSpace dashboard on the API Keys page.

How much does API generation cost?

Billing is per second and matches website pricing based on resolution (480p, 720p, or 1080p). Real-time token prices are returned by GET /v1/models and displayed on /v1/docs.