Skip to main content

Run Flow

POST 
/api/fluid-fit/run-flow

Executes a pre-configured flow by providing the appropriate input based on the flow's Initial Node(s).

Input Options:

  • imageFiles - Upload image files directly (maximum 10 files) — for flows with an Image Node entry
  • imageURLs - Array of image URLs (maximum 10 URLs) — for flows with an Image Node entry
  • videoFiles - Upload video files directly (maximum 10 files) — for flows with a Video Node entry
  • videoURLs - Array of video URLs (maximum 10 URLs) — for flows with a Video Node entry
  • text - A plain text string — for flows with a Text Node entry (e.g. Text-to-Audio flows)

A flow can have an Image root, a Video root, AND a Text root all feeding the same downstream node in parallel (e.g. an LLM node's separate Images/Videos/Text input anchors) — any combination of imageFiles/imageURLs, videoFiles/videoURLs, and text can be provided together in the same call; each is set on its matching Initial Node independently.

Flow Requirements:

  • An Initial Node (Image Node, Video Node, and/or Text Node)

Supported Inputs & Outputs:

InputFinal Node Output
imageFiles / imageURLsDepends on Final Node(s) — Image / Video / Audio / Text
videoFiles / videoURLsDepends on Final Node(s) — Image / Video / Audio / Text
textDepends on Final Node(s) — Image / Video / Audio / Text
Any combination of the aboveDepends on Final Node(s) — Image / Video / Audio / Text

The output type is always determined by the Final Node(s), not the input. A flow can produce any combination of Image, Video, Audio, or Text outputs regardless of what was passed in.

The input is set on the Initial Node(s), the flow is executed, and the processed output is returned from the Final Node(s).

Response: every item in result.outputs carries a type field — "text", "audio", "image", or "video" — making it easy to identify the type of each output. This is useful when a flow has multiple Final Nodes of different types (e.g. a generated image and a text summary in the same run).

Request​

Responses​

Flow executed successfully