fal.ai - BoostGPT Documentation

Any model hosted on fal.ai, called by id. Use it when a page needs a model the other connectors do not carry.

Connecting

In the dashboard, open Connectors, find fal.ai and press Connect. The key you enter is checked against fal.ai before anything is saved, so a key that is refused never becomes a connection that fails later.

What you will need

Field	What it is
API Key	Your fal.ai API key

Getting the credential

Create an API key at fal.ai/dashboard/keys
Paste it into the connector

Notes

Asynchronous. run_model returns a request_id. Poll get_status, then call get_result.
Finished media is copied onto BoostGPT storage, so the link keeps working.
Different models take different inputs. prompt fits nearly all of them; image_url covers image-to-image and image-to-video. A model needing something more unusual may not be reachable through these parameters.

What an agent can do with it

Tool	What it does
`run_model`	Start any model on fal.ai. `model` is the model’s path, e.g. ‘fal-ai/flux/dev’ for image generation or ‘fal-ai/kling-video/v2.5-turbo/pro/text-to-video’ for video. Put the model’s own inputs in the remaining parameters — `prompt` suits almost every model. Asynchronous: returns a request_id to poll with get_result.
`get_status`	Check whether a fal.ai job has finished. Status is ‘IN_QUEUE’, ‘IN_PROGRESS’ or ‘COMPLETED’. Call get_result once it is COMPLETED.
`get_result`	Collect the output of a completed fal.ai job. The finished image or video is rehosted on a durable URL. If the job has not finished this returns its status instead.
`cancel_request`	Cancel a fal.ai job that is still queued.

Useful resources