Video Generation API - BudgetPixel API

Text, images or video → video. Depending on the model: text-to-video, image-to-video from a start (and often an end) frame, reference-to-video, and editing or extending a video you supply. Many models also generate audio. Each model has its own endpoint, POST /v1/videos/{model}.

Models

19 models. Each name opens its API reference with every parameter.

Model	Maker	What it does
Kling 3.0 Standard `kling-v3.0-standard`	Kling	Kling 3.0 in Standard mode (720p): text-to-video, plus image-to-video with a start and an optional end frame.
Kling 3.0 Pro `kling-v3.0-pro`	Kling	Kling 3.0 in Professional mode (1080p): text-to-video, plus image-to-video with a start and an optional end frame.
Kling 3.0 4K `kling-v3.0-4k`	Kling	Kling 3.0 at native 4K: text-to-video, plus image-to-video with a start and an optional end frame.
Kling 3.0 Turbo `kling-3.0-turbo`	Kling	Fast text-to-video and image-to-video (start frame only) with native audio always included.
Kling V3 Omni `kling-v3-omni-video`	Kling	Text-to-video; image-to-video (start + optional end frame); reference-images-to-video (up to 7 images); and video-to-video editing (supply a `video`).
Kling v2.6 Pro `kling-v2.6-pro`	Kling	Text-to-video, plus image-to-video with a start frame and an optional end frame.
SeeDance 2.5 `seedance-2.5`	ByteDance	Flagship multimodal video: text-to-video, image-to-video with start and end frames, and reference-to-video from images, video clips and audio.
SeeDance 2.0 `seedance-2.0`	ByteDance	Text-to-video; image-to-video (start + optional end frame); reference-images-to-video (up to 9 images); and video-to-video editing (supply a `video`).
SeeDance 2.0 Fast `seedance-2.0-fast`	ByteDance	The same modes as SeeDance 2.0 (text / image / reference / video-to-video) at lower cost and latency.
SeeDance 2.0 Mini `seedance-2.0-mini`	ByteDance	The same modes as SeeDance 2.0 (text / image / reference / video-to-video) at the best cost-performance.
Wan 3.0 `wan-3.0-video`	Alibaba	All-in-one video generation up to 30 seconds at 30fps.
Wan 3.0 Prime `wan-3.0-video-prime`	Alibaba	The high-speed tier of Wan 3.0: the same text-, image- and reference-to-video modes, with much faster generation.
Wan 2.7 `wan-2.7-video`	Alibaba	Text-to-video; image-to-video with a start frame and an optional end frame; and video editing (supply a `video`).
HappyHorse 1.1 `happyhorse-1.1`	Alibaba	Highly realistic dynamic rendering: text-to-video, image-to-video from a start frame, and reference-images-to-video for subject consistency.
MiniMax H3 `minimax-h3`	MiniMax	Multimodal 2K video: text-to-video, image-to-video with start and end frames, and reference-to-video from images, video clips and audio.
MiniMax H3 Max `minimax-h3-max`	MiniMax	The post-trained H3 variant tuned for stronger prompt adherence and aesthetics, with native audio.
P-Video `p-video`	Pruna AI	Text-to-video; image-to-video with a start frame and an optional end frame; and audio-driven video from a supplied track.
P-Video 2 Pro `p-video-2-pro`	Pruna AI	Text-to-video, and image-to-video from a first frame, a last frame, or both.
FLUX 3 Video `flux-3-video`	Black Forest Labs	Text-to-video, image-to-video and video extension, with generated sound effects, ambience and lip-synced dialogue.

Modes, lengths and resolutions differ by model. Video generation modes compares them, and Video output format explains the codec each one returns.

How a request works

Send POST /v1/videos/{model} with your prompt and inputs. The response returns a job id right away.
Poll GET /v1/videos/{job_id} every 15 to 30 seconds until status is succeeded; video takes minutes. The response then holds the video URL.

Async jobs covers job statuses and errors, and Media inputs covers passing images, video and audio.

Pricing

Video is priced per second of output, by resolution, and some inputs add to that. GET /v1/models?type=video returns current prices, and POST /v1/cost quotes a request before you send it. You pay only for jobs that succeed; see Pricing & credits.