Digital Nirvana has launched RoughCut, a multi-user platform that applies AI to the logging and first-assembly stage of editing. Its distinguishing claim is that the system records an explanation for its choices alongside the generated cut, giving editors a trail they can inspect and revise.
During ingest, RoughCut detects scenes and shot types, transcribes speech, identifies faces and on-screen text, classifies audio and divides footage into usable clips. Teams can then guide an assembly through editorial instructions and continue the work in established finishing tools rather than accepting a locked output.
The platform is also intended to preserve what it learns about media so analysis can remain useful to an archive or media library. That could reduce repeated logging across newsroom, documentary and post-production work, while the written reasoning offers a governance aid for human review.
Digital Nirvana will show the product publicly for the first time at IBC2026. The announcement does not provide supported codec, MAM or NLE integrations, measured accuracy, security architecture or commercial availability, so the explainability and interoperability claims still require production evaluation.
