Create, edit, and automate video. The Rendley AI agent turns a prompt into an edited, rendered video; the same run can generate video, images, music and speech, transcribe, dub, lip sync and export MP4s.
Initial release. The Rendley AI Video Agent creates or edits a project from a prompt and optional files, then exports a video file (the Export a video file option, on by default).
Files are imported into the project through the Rendley API before the agent starts, so every attachment is complete when the agent reads it.
Projects are optional. Without one the Actor creates a project named from the prompt in the chosen workspace, picked by name or ID, with the first workspace as the default.
The run waits for the agent and the export and returns the video. It waits until just before its Apify timeout and then hands a still-running job off as running with its IDs instead of failing; a later run with the ID in Job ID picks it up. The export is started once per agent job and remembered, so repeated pick-ups never export twice.
The exported video is streamed into the key-value store rather than held in memory, so the Actor runs in 128 MB, the smallest memory Apify offers, by default.
Results follow the Rendley API. The dataset item carries url, media_id, file_hash, result_data, last_message, commands_applied and the raw agent_job and export_job, and the video is copied to the run's key-value store as OUTPUT_VIDEO.
