Media hours & AI Credits

RiverScript uses two separate usage systems depending on the type of task:

• Media hours — used for transcription. One hour of audio or video content costs one media hour.
• AI credits — used for translation, AI Summarize, and AI Ask (chat with transcript).

Usage is charged per request, not per result

Media hours and AI credits are consumed when a request is sent to an AI model — not based on the quality or usefulness of the result. Every model call incurs a cost regardless of outcome, and usage is measured accordingly.

This means:

• If you transcribe a one-hour file, one media hour is consumed regardless of whether the transcript meets your expectations.
• If you translate a transcript, the corresponding AI credits are consumed regardless of whether the result suits your needs.
• If a file appears valid to our system and is sent to the transcription model, media hours are consumed — even if the audio turns out to be silent or unrecognizable.

If a file is rejected before reaching the model (for example, due to an unsupported format or a failed upload), no hours or credits are charged.

Why you cannot stop an in-progress transcription

RiverScript has a Stop Queue button rather than a simple stop button. This distinction is intentional.

Once a file has been sent to the transcription model, processing cannot be interrupted — the model call has already been made and will run to completion. Media hours for that file will be consumed regardless.

Stop Queue means: do not send any further files to the model. The file currently being processed will finish and its hours will be charged. Files waiting in the queue will not be sent.

This is why we recommend reviewing your queue before starting transcription, especially when processing multiple files at once.