rocketkind
AI

fal.ai

fal.ai is a hosted inference platform for generative media: image, video, audio and text models (FLUX, Kling, Whisper, LLM routers and hundreds more), each exposed as its own endpoint and billed per use.

Authentication

  • API key

Actions

Generic API request. Call any fal Platform API endpoint on this connection - assets, storage, workflows, serverless apps, API keys, billing events - for the case where no shipped action covers it yet. The connection supplies the base URL, the Authorization: Key header and the egress route, so only the path, query and body are yours to…

  • Cancel request (v1)DestructiveAsk fal to cancel a queued or running request, to stop paying for a generation you no longer need.
  • Estimate cost (v1)Estimate what a planned volume of fal usage will cost across one or more models, before committing to a batch.
  • Get account billing (v1)Read the fal account's username and, with Include credits set, its remaining credit balance - to alert before credits run out or to gate a large batch on having enough left.
  • Get model (v1)Look up one fal model endpoint including its full OpenAPI 3.0 document - the way to learn what a model accepts before building the input object for Submit request or Run model, and what its output looks like.
  • Get model pricing (v1)Look up what fal charges for one or more model endpoints before running them - to pick the cheaper of two models, or to budget a batch.
  • Get request result (v1)Fetch the output of a fal request once Get request status reports COMPLETED - the generated images, video, audio or text, exactly as the model returns it (media arrives as fal.media URLs, not bytes).
  • Get request status (v1)Check where a queued fal request stands - poll this after Submit request until body.status is COMPLETED, then fetch the output with Get request result.
  • Get usage (v1)Report what this fal account consumed and spent over a time window, bucketed by minute, hour, day, week or month - for cost dashboards, budget alerts or charge-back per model or API key.
  • List models (v1)Search fal's model gallery to find the endpoint id to run - by free text, category (text-to- image, image-to-video, ...) or status.
  • List requests (v1)List the requests this account made to a model endpoint, newest first, to find a request id you did not keep, audit failures, or read what was sent and returned.
  • Run model (v1)DestructiveRun a fal model endpoint and wait for its output in the same call - for fast models (a few seconds, such as fal-ai/flux/schnell or an LLM endpoint) where queueing and polling is overhead.
  • Submit request (v1)Queue a run of any fal model endpoint and return immediately with a request id - the default way to run a fal model, because generation (images, video, audio) often outlives a single HTTP call.

Product names, logos and brands are the property of their respective owners and are used here for identification only. Their use does not imply endorsement by, or affiliation with, those owners.