Is there one API for Seedance, Kling and Veo?

Yes. varg runs Seedance 2.5 ("seedance_2_5"), Kling 4.0 ("kling_v4"), Veo 3.1 ("veo_3_1") and 80 more video models behind one endpoint, POST https://api.varg.ai/v2/tools/video/call, with one API key and one credit balance. The request has the same shape for each: change the "model" string to switch.

How much does the AI video API cost?

Every model is billed at its provider's price with no markup, from one credit balance: 1 credit is $0.01. For example: Seedance 2.5 from $0.43 per run, Kling 4.0 from $0.48 per run, Veo 3.1 from $0.80 per run, Wan 3.0 from $0.10 per run and FLUX 3 from $0.85 per run. Among the video models with a stated run price, the lowest is Creatify Boreal: from $0.01 per run. For the exact price of a request before you run it, send the same body to POST https://api.varg.ai/v2/estimate.

Which AI video models take an image as input?

51 of the 83 have a route that starts from an image, among them Seedance 2.5, Kling 4.0, Veo 3.1, Wan 3.0, FLUX 3, MiniMax H3 and 8 more. Pass the image as a URL in "files".

Which AI video models generate audio with the video?

20 of the 83 take "audio": true in the request and generate the video with a native audio track, among them Seedance 2.5, Kling 4.0, Veo 3.1, FLUX 3, LTX-2, Kandinsky 6.0 Pro and 1 more. The input schemas of the others state no audio switch; each model page says what that model returns.

Which AI video model makes the longest clip?

By the models' input schemas the longest request is 30 seconds, on Kling 4.0, Seedance 2.5 and Wan 3.0.

How do I call the AI video API?

Send POST https://api.varg.ai/v2/tools/video/call with the header "Authorization: Bearer $VARG_API_KEY" and a JSON body that includes "model", for example "model": "seedance_2_5". The call answers 202 with a job right away. Poll GET https://api.varg.ai/v2/jobs/{id} until "status" is "completed", then read output.outputs[0].url, or pass options.webhook_url to be called back.

Can I use these AI video models from Claude, ChatGPT or Cursor?

Yes. Add the varg MCP server (https://mcp.varg.ai/mcp) to Claude, ChatGPT, Cursor or any other MCP client and sign in with varg, with no API key to paste. The agent runs a model with call_tool, passing tool "video" and the model id, for example "model": "seedance_2_5".

83 models · 19 families · one API key

AI video API

varg runs 83 AI video models behind one endpoint and one API key. 19 model families are listed below, with their flagships: Seedance 2.5, Kling 4.0, Veo 3.1, Wan 3.0 and 15 more. The request has the same shape for each of them: the "model" string picks the model.

Every model is billed at its provider's price with no markup, from one credit balance: 1 credit is $0.01.

POST /v2/tools/video/call

curl -X POST https://api.varg.ai/v2/tools/video/call \
  -H "Authorization: Bearer $VARG_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "seedance_2_5",
  "prompt": "A cyberpunk city street at night, neon signs reflecting in puddles. A heavy-set woman in a worn military uniform stands breathing hard, out of shape. She begins military calisthenics — push-ups, burpees, mountain climbers — her body transforming in real-time: fat melting away, muscles defining, posture straightening. Holographic HUD overlays track her progress with glowing stats. Rain falls around her as she powers through the workout, her uniform tightening as she becomes lean and athletic. The transformation completes — she stands as a fit soldier-athlete, rain-soaked, neon-lit, triumphant. Cinematic, dramatic lighting, cyberpunk aesthetic.",
  "aspect_ratio": "16:9",
  "resolution": "720p",
  "duration": 5
}'

To run another model, change "model": kling_v4, veo_3_1, wan_3, flux_3, minimax_h3.

Try Seedance 2.5 in the Playground Get an API key

AI video model families

Each family has a page with all of its models, and each model a page with its parameters. The price is the flagship's own, in US dollars, with no markup.

AI video model families on varg
Family	Flagship model	Catalog description	Flagship price	Models
Seedance API	Seedance 2.5`seedance_2_5`	ByteDance Seedance 2.5 — the flagship Seedance model. High-quality video from text, with optional image (≤9), video (≤3), and audio (≤3) references…	From $0.43 per runfrom $0.1075/s	7
Kling API	Kling 4.0`kling_v4`	Kling 4.0 by Kuaishou. Up to 30 seconds at 1080p, multi-shot prompts, up to 7 reference elements, 10 keyframes and video editing.	From $0.48 per runfrom $0.157/s	14
Veo API	Veo 3.1`veo_3_1`	Google Veo 3.1. Text-to-video and image-to-video with native audio, including first and last frame control, up to 4K.	From $0.80 per runfrom $0.20/s	3
Wan API	Wan 3.0`wan_3`	Alibaba Wan 3.0. Top-ranked text-to-video — up to 30s at 1080p with native audio, first/last-frame animation, and references (≤10 images, ≤5 videos…	From $0.10 per runfrom $0.05/s	3
FLUX API	FLUX 3`flux_3`	Black Forest Labs FLUX 3. Latest flagship video model — text-to-video, image-to-video, and video-to-video with native audio. 720p/1080p, 5-10s.	From $0.85 per runfrom $0.17/s	3
MiniMax API	MiniMax H3`minimax_h3`	MiniMax H3. Frontier 2K video model with native stereo audio. Text, image, or multimodal reference inputs.	From $0.25 per runfrom $0.05/s	17
HeyGen API	HeyGen Avatar IV`heygen_avatar_iv`	HeyGen Interactive Avatar. Next-gen avatar with improved visual fidelity.	$0.0667/s	5
Grok Imagine API	Grok Imagine 1.5`grok_imagine_1_5`	Grok Imagine Video 1.5 by xAI. Text, image and multi-reference video with native audio, 1-15 seconds up to 1080p.	From $0.08 per runfrom $0.08/s	4
Gemini Omni API	Omni Flash 1.2`omni_flash_1_2`	Google Gemini Omni Flash 1.2 on fal. It makes 3-15 s video with native audio from text, from a first/last frame, or from up to 10 reference images…	From $0.0273/s	3
OmniHuman API	OmniHuman v2`omnihuman_v2`	OmniHuman v2 by ByteDance. A talking video from one image and an audio track.	From $2.80 per run	2
PixVerse API	PixVerse V6 Max`pixverse_v6_max`	PixVerse V6 Max: 1-15 s video from text or a first-frame image at 540p/720p/1080p (360p via provider_options), with optional multi-clip camera…	From $0.081/s	2
LTX API	LTX-2`ltx_2_19b_distilled`	Lightricks LTX-2 distilled.	From $0.50 per run	2
Kandinsky API	Kandinsky 6.0 Pro`kandinsky_6`	Kandinsky 6.0 Pro by Kandinsky Lab. Open-weight text- and image-to-video with a synchronized audio track; 5-second clips, upscalable to 1080p or 4x.	From $0.24 per run$0.03/video	2
Vidu API	Vidu Q4`vidu_q4`	Vidu Q4. Video with native audio from a first frame, or from reference images and voice clips; 3-16 seconds, 540p to 4K. No text-only mode.	From $0.14 per runfrom $0.045/s	1
Happy Horse API	Happy Horse`happy_horse`	Alibaba video generation. Expressive character motion.	From $0.42 per runfrom $0.14/s	1
Sync lipsync API	Sync v3`sync_v3`	Sync Labs v3.	From $1.00 per run	3
ID-V2V API	ID-V2V`id_v2v`	ID-V2V (Eyeline Labs / Netflix research). Restyle a video's scene, lighting and look from an edited first frame while keeping the subjects'…	From $0.20 per run$0.20/s	2
Topaz API	Topaz Video Upscale`topaz_video`	Topaz-quality video upscaling.	From $0.50 per run	1
SeedVR API	SeedVR Video Upscale`seedvr_video`	ByteDance SeedVR video upscaler. Good quality video enhancement.	From $0.30 per run	1

Call the AI video API

One endpoint for every video model. The call below runs Seedance 2.5; the "model" string picks the model, and the rest of the request keeps its shape.

REST

POST https://api.varg.ai/v2/tools/video/call with Authorization: Bearer $VARG_API_KEY. It answers 202 with a job right away; poll GET /v2/jobs/{id} until its status is completed, then read output.outputs[0].url.

POST /v2/tools/video/call

curl -X POST https://api.varg.ai/v2/tools/video/call \
  -H "Authorization: Bearer $VARG_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "seedance_2_5",
  "prompt": "A cyberpunk city street at night, neon signs reflecting in puddles. A heavy-set woman in a worn military uniform stands breathing hard, out of shape. She begins military calisthenics — push-ups, burpees, mountain climbers — her body transforming in real-time: fat melting away, muscles defining, posture straightening. Holographic HUD overlays track her progress with glowing stats. Rain falls around her as she powers through the workout, her uniform tightening as she becomes lean and athletic. The transformation completes — she stands as a fit soldier-athlete, rain-soaked, neon-lit, triumphant. Cinematic, dramatic lighting, cyberpunk aesthetic.",
  "aspect_ratio": "16:9",
  "resolution": "720p",
  "duration": 5
}'

From Claude, ChatGPT or Cursor (MCP)

The varg MCP server, https://mcp.varg.ai/mcp, runs Seedance 2.5 with call_tool. Sign in with varg, no API key to paste. How to connect it

Claude Code

claude mcp add --transport http varg https://mcp.varg.ai/mcp

call_tool

{
  "tool": "video",
  "input": {
    "model": "seedance_2_5",
    "prompt": "A cyberpunk city street at night, neon signs reflecting in puddles. A heavy-set woman in a worn military uniform stands breathing hard, out of shape. She begins military calisthenics — push-ups, burpees, mountain climbers — her body transforming in real-time: fat melting away, muscles defining, posture straightening. Holographic HUD overlays track her progress with glowing stats. Rain falls around her as she powers through the workout, her uniform tightening as she becomes lean and athletic. The transformation completes — she stands as a fit soldier-athlete, rain-soaked, neon-lit, triumphant. Cinematic, dramatic lighting, cyberpunk aesthetic.",
    "aspect_ratio": "16:9",
    "resolution": "720p",
    "duration": 5
  },
  "wait": true
}

From a terminal (CLI)

vargai is the varg command line: every tool of the MCP server is a command. Install and log in

vargai

curl -fsSL https://varg.ai/install | sh
vargai call_tool --tool video '{"model":"seedance_2_5","prompt":"A cyberpunk city street at night, neon signs reflecting in puddles. A heavy-set woman in a worn military uniform stands breathing hard, out of shape. She begins military calisthenics — push-ups, burpees, mountain climbers — her body transforming in real-time: fat melting away, muscles defining, posture straightening. Holographic HUD overlays track her progress with glowing stats. Rain falls around her as she powers through the workout, her uniform tightening as she becomes lean and athletic. The transformation completes — she stands as a fit soldier-athlete, rain-soaked, neon-lit, triumphant. Cinematic, dramatic lighting, cyberpunk aesthetic.","aspect_ratio":"16:9","resolution":"720p","duration":5}'

How billing works

Every model is billed at its provider's price with no markup, from one credit balance: 1 credit is $0.01. varg's fee is on the top-up, not on the run: 5.5% of the credits you buy ($0.80 at the least), or 8% for a team workspace. Pricing

AI video models compared

Price, limits and inputs side by side, from the same catalog as this page.

How to use it

Step-by-step pages on varg how to.

AI video API: questions

Is there one API for Seedance, Kling and Veo?: Yes. varg runs Seedance 2.5 ("seedance_2_5"), Kling 4.0 ("kling_v4"), Veo 3.1 ("veo_3_1") and 80 more video models behind one endpoint, POST https://api.varg.ai/v2/tools/video/call, with one API key and one credit balance. The request has the same shape for each: change the "model" string to switch.
How much does the AI video API cost?: Every model is billed at its provider's price with no markup, from one credit balance: 1 credit is $0.01. For example: Seedance 2.5 from $0.43 per run, Kling 4.0 from $0.48 per run, Veo 3.1 from $0.80 per run, Wan 3.0 from $0.10 per run and FLUX 3 from $0.85 per run. Among the video models with a stated run price, the lowest is Creatify Boreal: from $0.01 per run. For the exact price of a request before you run it, send the same body to POST https://api.varg.ai/v2/estimate.
Which AI video models take an image as input?: 51 of the 83 have a route that starts from an image, among them Seedance 2.5, Kling 4.0, Veo 3.1, Wan 3.0, FLUX 3, MiniMax H3 and 8 more. Pass the image as a URL in "files".
Which AI video models generate audio with the video?: 20 of the 83 take "audio": true in the request and generate the video with a native audio track, among them Seedance 2.5, Kling 4.0, Veo 3.1, FLUX 3, LTX-2, Kandinsky 6.0 Pro and 1 more. The input schemas of the others state no audio switch; each model page says what that model returns.
Which AI video model makes the longest clip?: By the models' input schemas the longest request is 30 seconds, on Kling 4.0, Seedance 2.5 and Wan 3.0.
How do I call the AI video API?: Send POST https://api.varg.ai/v2/tools/video/call with the header "Authorization: Bearer $VARG_API_KEY" and a JSON body that includes "model", for example "model": "seedance_2_5". The call answers 202 with a job right away. Poll GET https://api.varg.ai/v2/jobs/{id} until "status" is "completed", then read output.outputs[0].url, or pass options.webhook_url to be called back.
Can I use these AI video models from Claude, ChatGPT or Cursor?: Yes. Add the varg MCP server (https://mcp.varg.ai/mcp) to Claude, ChatGPT, Cursor or any other MCP client and sign in with varg, with no API key to paste. The agent runs a model with call_tool, passing tool "video" and the model id, for example "model": "seedance_2_5".

For AI agents: every model has its full spec as Markdown, listed in varg.ai/models/llms.txt.