AI-assisted audio description
Audio description for video that was never described.
SceneTell creates timed narration tracks that make visual media easier to follow: action, setting, on-screen text, gestures, scene changes, and the details that carry the story.
The product is built around the hard part of audio description: saying enough, at the right moment, without burying the original soundtrack.
Rights-safe demos are being prepared.
Public examples will use media we can publish and a SceneTell-controlled accessible HTML5 player.
Built around timing, not just text generation.
Effective description depends on what the video shows, what the soundtrack is already saying, and how much space exists between them. SceneTell treats that as the product, not an afterthought.
-
Analyze the source
Transcript timing, speech activity, loudness, scene changes, and on-screen text help SceneTell understand where description can fit.
-
Develop visual context
Scene structure, character context, slice-level narration, and OCR evidence help the system track what matters across the video.
-
Render reviewable outputs
Drafted narration becomes a script, a narration track, mixed audio, and video deliverables that can be listened to, evaluated, and improved.
For individuals now. For organizations next.
The first use case is direct: a blind viewer has a video and needs it described. The same problem exists at organizational scale in training libraries, museum media, education, product content, and internal communications.
-
Personal access
For videos people encounter before a platform, creator, or studio has provided description.
-
Media libraries
For teams that need a practical path through existing video backlogs and future publishing workflows.
-
Deliberate rollout
Launch starts with a waitlist so we can onboard real use cases while scaling processing capacity and product quality.
Join the SceneTell waitlist.
SceneTell is entering a controlled launch. Join the waitlist to be considered for early access as capacity opens and we learn from real videos, real workflows, and real users.