VidFlux API & MCP

Generate AI videos and images from your own code, or let AI agents like Claude and Cursor do it through MCP. Every generation is billed to your VidFlux credits at the same price as the website, and failed generations are refunded automatically.

1. Get an API key

Authorization: Bearer vf_your_api_key

2. Create a video

curl -X POST https://vidflux.ai/api/v1/videos \
  -H "Authorization: Bearer vf_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedance-2",
    "prompt": "A corgi surfing a wave at sunset, cinematic",
    "aspect_ratio": "16:9",
    "duration": 5
  }'

Response:

{
  "id": "3f1c9e0a-...",
  "type": "video",
  "model": "seedance-2",
  "status": "processing",
  "credits_used": 425
}

For image-to-video, pass public image URLs in image_urls. Add "dry_run": true to any create request to get the exact credit price without generating.

3. Poll the task

Generation is asynchronous and usually takes 1 to 5 minutes. Poll every 10 to 20 seconds until status is completed or failed.

curl https://vidflux.ai/api/v1/tasks/3f1c9e0a-... \
  -H "Authorization: Bearer vf_your_api_key"

{
  "id": "3f1c9e0a-...",
  "type": "video",
  "status": "completed",
  "video_url": "https://...mp4",
  "error": null,
  "credits_used": 425
}

Create an image

curl -X POST https://vidflux.ai/api/v1/images \
  -H "Authorization: Bearer vf_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{"model":"nano-banana-pro","prompt":"Product photo of a glass perfume bottle","resolution":"2K"}'

Poll /api/v1/tasks/{id} the same way; the result is in image_url.

Endpoints

POST /api/v1/videos	Create a video task
POST /api/v1/images	Create an image task
GET /api/v1/tasks/{id}	Task status and result URL
GET /api/v1/models	Models, parameters (JSON Schema) and default prices
GET /api/v1/credits	Your credit balance

Errors return {"error":{"code","message"}} with status 400 (invalid parameters), 401 (bad key), 402 (insufficient credits), 404 or 502 (provider failure).

Models

Model id	Description	Credits (defaults)
seedance-2	Seedance 2.0. Text-to-video or image-to-video (up to 2 frame images, or reference images with use_reference_images). 4-15s, 480p/720p, optional audio.	680
seedance-2-fast	Seedance 2.0 Fast. Text-to-video or image-to-video (up to 2 frame images, or reference images with use_reference_images). 4-15s, 480p/720p, optional audio.	544
seedance-2-mini	Seedance 2.0 Mini. Text-to-video or image-to-video (up to 2 frame images, or reference images with use_reference_images). 4-15s, 480p/720p, optional audio.	360
seedance-2.5	Seedance 2.5. Text-to-video or image-to-video (up to 2 frame images, or reference images with use_reference_images). 4-30s, 480p/720p, optional audio.	704
seedance-1-pro	Seedance 1.0 Pro. 5 or 10s at 480p/720p/1080p, text-to-video or image-to-video with optional end frame.	100
kling-3	Kling 3.0 Standard. Text-to-video or image-to-video, 3-15s, optional sound and multi-shot.	200
kling-3-pro	Kling 3.0 Pro. Text-to-video or image-to-video, 3-15s, optional sound and multi-shot.	300
kling-2.6	Kling 2.6. 5 or 10s, text-to-video or image-to-video, optional sound.	150
kling-2.5	Kling 2.5 Turbo. 5 or 10s, text-to-video or image-to-video, negative prompt supported.	110
kling-3-motion	Kling 3.0 Motion Control. Transfers the motion of a reference video (video_url) onto the character in image_url. Billed per second of video_duration (the reference video length).	200
kling-2.6-motion	Kling 2.6 Motion Control. Transfers the motion of a reference video (video_url) onto the character in image_url. Billed per second of video_duration (the reference video length).	125
veo-3.1	Veo 3.1 Quality. 8s video with native audio, text-to-video or image-to-video (1 start frame, or 2 for start and end frames).	280
veo-3.1-fast	Veo 3.1 Fast. 8s video with native audio, text-to-video or image-to-video (1 start frame, or 2 for start and end frames).	120
minimax-h3	MiniMax Hailuo H3. 4-15s at 720p or 2K. Up to 2 frame images, or up to 9 reference images with use_reference_images.	175
gemini-omni-video	Gemini Omni Video. 4/6/8/10s at 720p, 1080p or 4K, up to 7 reference images.	200
grok-imagine	Grok Imagine. 6s video, fun/normal/spicy modes, up to 7 images. Pro members only.	96
gpt-image-2.5	GPT Image 2.5. Text-to-image, or editing with up to 16 reference images, 1K/2K/4K. Strong text rendering.	20
gpt-image-2	GPT Image 2. Text-to-image, or editing with up to 16 reference images, 1K/2K/4K. Strong text rendering.	20
nano-banana-pro	Nano Banana Pro (Gemini 3 Pro Image). Text-to-image or editing with up to 10 reference images, 1K/2K/4K.	100
nano-banana-2.1	Nano Banana 2.1. Text-to-image or editing with up to 10 reference images, 1K/2K/4K.	15
nano-banana-2	Nano Banana 2. Text-to-image or editing with up to 14 reference images, 1K/2K/4K.	30

Exact parameters per model are returned by GET /api/v1/models.

MCP server

The MCP endpoint https://vidflux.ai/api/mcp (Streamable HTTP) exposes the tools generate_video, generate_image, get_task, list_models and get_credits, using the same API key. See the MCP server guide for setup in every client.

Claude Code

claude mcp add --transport http vidflux https://vidflux.ai/api/mcp \
  --header "Authorization: Bearer vf_your_api_key"

Cursor, Windsurf, VS Code

{
  "mcpServers": {
    "vidflux": {
      "url": "https://vidflux.ai/api/mcp",
      "headers": { "Authorization": "Bearer vf_your_api_key" }
    }
  }
}

Claude Desktop

{
  "mcpServers": {
    "vidflux": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://vidflux.ai/api/mcp",
               "--header", "Authorization: Bearer vf_your_api_key"]
    }
  }
}