Black Forest Labs: FLUX Video Edit

FLUX Video Edit is a video model from Black Forest Labs, the same provider behind the FLUX family of image models. It sits in the video category on api14 and is identified by the model id black-forest-labs/flux-video-edit. The provider describes it as a fast editing model rather than a generator from scratch: it takes an existing clip and an instruction, then returns a new version of that clip with the requested change applied.

The model accepts two input modalities, text and video, and produces video as its only output. The text side is an edit prompt, and the video side is the source clip you want changed. According to the provider description, the edits it can perform include adding, removing, or replacing objects and characters, rebuilding the setting, and editing on-screen elements. It does not generate audio, so any soundtrack or dialogue in your source needs to be handled outside the model. The facts provided do not include a context window, resolution options, aspect ratios, or duration limits, so treat those as unspecified and check the live model page for current details.

Because the model works by transforming a supplied clip, the practical use cases are editorial. You can take a shot and remove an object that should not be there, swap one character or prop for another, or replace a background to move a scene into a different setting. It also fits workflows where on-screen content needs to change without reshooting, such as updating signage or other visible elements in a clip. If your goal is to generate a video from a text description alone, this is not the right tool, since it expects a source video as part of the request.

You call it through api14's OpenAI-compatible API using the model id black-forest-labs/flux-video-edit. The request carries your edit prompt as text and the source video as the video input, and the response returns the edited video. The supported parameters list for this model is empty, so there are no documented tuning options such as resolution, aspect ratio, duration, or frame controls to set on the request. Audio generation is off. In practice that means you send the prompt and the source clip and take the model's default output, rather than steering the render with extra fields.

Given the empty parameter list and the absence of documented resolution, aspect ratio, and duration values, this model is best treated as a fixed-configuration editing endpoint. If you need fine control over output dimensions, clip length, or frame handling, or if you need a model that generates video from text without a source clip, another model on api14 may fit better. For straightforward instruction-based edits to an existing video, it covers the core cases the provider describes.