AI video generation is only one part of a delivery workflow. Teams may already have a generated clip, product video, campaign asset or archive file that needs a cleaner high-resolution output.
A standalone Video Enhancement API processes that existing source without asking the generation model to create a new scene. Through Light Speed Future, developers can submit a source video, choose an enabled enhancement route and retrieve the completed result through the same asynchronous video-task pattern used across the LSF platform.
When to use standalone enhancement
Use standalone Video Enhancement when:
- The source video already exists.
- You want enhancement without generating a new scene.
- The application needs explicit resolution, FPS or bitrate controls.
- You need to compare Standard and Professional processing.
- Enhancement should remain a separate step in your product interface.
If the application needs Seedance generation and enhancement to happen automatically in one task, use LSF Seedance2.5 Production instead.
Lite, Standard and Professional routes
LSF exposes three standalone processing routes.
Lite
Lite uses a dedicated model alias. It supports the common resolution, FPS and bitrate controls for a streamlined enhancement workflow. It does not use tool_version="fast" and does not accept Standard- or Professional-only controls.
Standard
Standard uses the shared Standard/Professional alias with tool_version omitted or set to standard. It supports scene and enhancement-style controls.
Professional
Professional uses the same shared alias with tool_version="professional". It supports professional enhancement style and established output-depth controls.
Lite supports outputs through 4K. Standard and Professional support outputs through 8K when enabled, covering 14 tier/resolution price bands in total. The enabled resolutions and advanced media options are product- and customer-specific. Do not assume every combination is available until it has been confirmed.
Core request fields
A standalone enhancement request uses POST /v1/videos and supplies an existing source through metadata.video_url. The customer also supplies the source duration for the initial reservation.
Common controls include:
metadata.resolutionfor an enabled named output such as 1080p, 2K or 4K.metadata.fpsfor a target between 15 and 120 FPS.metadata.bitrate_levelfor low, medium or high.metadata.bitratefor an explicit supported bitrate.metadata.scenefor Standard scene treatment.metadata.enhance_stylefor Standard or Professional.metadata.bit_depthfor supported Professional output.metadata.client_request_idfor safe identical-request replay.
The source URL must be a direct public HTTP or HTTPS address that remains accessible during validation and processing. A URL that plays in a browser may still fail if it requires cookies, private headers, unstable redirects or anti-bot access.
Resolution and FPS affect the final cost
Standalone enhancement is billed from the actual enhanced output duration, selected output profile and FPS band. The output duration is rounded up to a whole second.
| Actual output FPS | Multiplier |
|---|---|
| Up to 30 FPS | ×1 |
| More than 30, up to 60 FPS | ×2 |
| More than 60, up to 120 FPS | ×4 |
If the request omits FPS, Light Speed Future reserves at the highest band while leaving FPS omitted in the upstream enhancement request. Final billing then uses the actual output FPS. Omitting FPS does not force the video to 120 FPS.
This distinction should be visible near any standalone enhancement price table. A base 1080p price without its FPS condition is incomplete.
Design for asynchronous delivery
An accepted create response means that the job entered processing; it does not mean the final video is ready. Store the returned task ID and retrieve the same task through GET /v1/videos/{task_id} until it reaches a terminal state.
For reliable integration:
- Validate the source before submitting it.
- Store the exact request and
client_request_id. - Poll the original task every 30 seconds or longer.
- Do not create another task because a status request was interrupted.
- Read the actual output duration, resolution and FPS from the completed result.
- Copy the temporary output into customer-controlled storage.
Standalone enhancement vs Production
Standalone enhancement expects a source video and returns an enhanced version of that source. Production starts with a prompt and supported references, generates a new Seedance 2.5 video, enhances it and returns the finished result through one task.
Their billing models also differ. Standalone enhancement uses an FPS multiplier; Production currently does not. Production may count measured reference-video duration together with rounded output duration, while standalone enhancement bills the enhanced output under its own contract.
Add video enhancement to your product
Use standalone LSF Video Enhancement when your users bring an existing video. Use LSF Seedance2.5 Production when they want to create and finish a new video through one managed workflow.
View current pricing or contact Light Speed Future to confirm enabled routes and output combinations.