Use this in the request body model field.
Video APIActivekling-3 API
Generate AI videos with the Kling 3.0 API via API in One. Supports text-to-video and image-to-video generation with dynamic credits based on mode, duration, and sound.
The currently implemented BFF route.
Input types: text, image
Keys are managed by the shared dashboard.
Use callback_url on supported async tasks.
Live capabilities
These capabilities come from the current model config and backend route. Unreleased APIs are not listed.
Transparent dynamic credit pricing by mode
Transparent dynamic credit pricing by mode, duration, and sound
Both text
Both text-to-video and image-to-video in one endpoint
Multi
Multi-shot mode for multi-scene videos
Sound generation support (audio in video)
Sound generation support (audio in video)
Request format
Clients call same-origin API routes; the server BFF forwards to the matching Worker.
MethodEndpointCreditsPurpose
POST/api/v1/videos/generations300 creditsState-of-the-art AI video generation with 4K quality support. Text-to-video and image-to-video capabilities.
Parameters
promptstringText description of the video (max 2500 chars). Not required if multi_shots is true.
RequiredmodestringQuality mode: "std" (standard) or "pro" (professional)
Optional · Default stddurationnumberVideo duration in seconds (3–15)
Optional · Default 5aspect_ratiostringOutput aspect ratio: 1:1, 9:16, 16:9
Optional · Default 16:9soundbooleanEnable sound/audio generation in the video
Optional · Default trueimagestringImage URL for image-to-video mode
Optionalmulti_shotsbooleanEnable multi-shot mode for multi-scene videos
Optional · Default falsemulti_promptarrayArray of {prompt, duration} objects for multi-shot mode
Optionalkling_elementsarrayKling-specific creative elements
OptionalExample request
curl -X POST https://apiin.one/api/v1/videos/generations \
-H "Authorization: Bearer aio_your_api_key" \
-H "Content-Type: application/json" \
--data-raw '{
"model": "kling-3",
"prompt": "A golden retriever running through a field of sunflowers at sunset, cinematic quality",
"mode": "std",
"duration": 5,
"aspect_ratio": "16:9",
"sound": true
}'Models on the same endpoint
These live models use the same implemented BFF endpoint and can be switched by changing the model field.
grok-videovideo · 24 credits/api/v1/videos/generations🎞️ltx-2video · 24 credits/api/v1/videos/generations
seedance-1-5-provideo · 28 credits/api/v1/videos/generations
hailuo-2-3video · 30 credits/api/v1/videos/generations
wan-2-6video · 80 credits/api/v1/videos/generations
pixverse-5-6video · 90 credits/api/v1/videos/generationsIntegration flow
These steps use the current live route and shared account system; login, payment, orders, credits, and profile remain shared.
Create a shared API key
Create a key in the existing dashboard. Top-up, orders, credit balance, and profile stay in the shared console.
Call the real endpoint
Send a request with the endpoint, model id, and parameters shown on this page. /api/v1/videos/generations
Poll task status
Async media and tool jobs use /api/v1/tasks/{task_id}; chat models return a direct response.
Task status route
After creating image, video, audio, or tool tasks, poll the real task endpoint for results.
GET/api/v1/tasks/{task_id}Check generation task status and result.Pricing
Dynamic: 120–1200 credits per video (~$1.20–$12.00); default std 5s with audio is 300 credits (~$3.00)
Base cost$3.00
Credits300
BillingSuccessful requests
Use cases
These workflows are supported by the current model and backend node.
Create marketing and social media video content from text prompts
Animate product photos into dynamic video ads
Build AI video generation features into your SaaS product
Generate cinematic video clips for filmmaking and creative projects
FAQ
How much does the Kling 3.0 API cost?
Pricing depends on mode (std/pro), duration (3–15s), and sound. Default std 5s with audio costs 300 credits. The pricing formula is: std no-audio 40/s, std with-audio 60/s, pro no-audio 54/s, pro with-audio 80/s.
What is the difference between Pro and Standard mode?
Pro mode produces higher quality videos with better motion and detail. Standard mode is faster and costs fewer credits. Specify via the "mode" parameter (std or pro).
Can I generate videos from images?
Yes. Provide image URLs in the request and the model auto-detects image-to-video mode. You don't need to set the task type manually.
What video durations are supported?
Kling 3.0 supports 3–15 second video generation. Set the "duration" parameter accordingly. Default is 5 seconds.
What is multi-shot mode?
Multi-shot mode lets you create videos with multiple scenes, each with its own prompt and duration. Set multi_shots to true and provide a multi_prompt array.