AirCaption: Fast Audio & Video Transcription & Subtitles
AirCaption is a Mac and Windows application for transcribing audio and video and preparing captions. Its workflow combines speech recognition with text and timing edits, optional translation, and exports. The publisher names NVIDIA, Cohere, and Whisper models running on your computer.
That local workflow is useful when uploading a large recording to a transcription website would be inconvenient, or when you want to edit captions alongside the media. It does not remove the need to check the resulting words and timing.
Transcribe, then correct the captions
Generate captions from the recording, then review them with the application's waveform, hotkeys, and find-and-replace tools. AirCaption also imports existing caption files, so the starting point does not have to be a new automatic transcript.
Check names, specialist terminology, and boundaries between speakers before delivery. For subtitles, a correct sentence with badly aligned timing can still be difficult to follow. Use the editing stage to fix both, particularly if the output will accompany a course, interview, or public video.
Exports include SRT, VTT, plain text, and video with captions burned in. Choose a subtitle file when the next platform accepts a separate caption track; a plain transcript and a visibly captioned video serve different purposes. Confirm the receiving platform's requirements before exporting.
Distinguish local processing from cloud translation
The publisher says the application runs locally by default and uploads nothing unless you choose to. Optional cloud translation sends the selected caption text, rather than being part of an entirely offline promise.
The site advertises subtitle translation in more than 80 languages. Check the language and processing option you intend to use, since local and cloud arrangements can differ. The available description is a publisher statement about processing, not an independent security assessment or a guarantee that every model suits your hardware.
Folder transcription is also advertised, but adding multiple files to the transcription queue is specifically listed under Pro. Do not assume that the homepage's batch example is included with every plan.
Match the plan to models and translation volume
The pricing page lists Essentials as free, with unlimited AI transcription, editing, caption exports, and ten cloud-translation minutes per month. Pro is $9.99 per month in US dollars and adds medium and large models, multi-file queuing, and 30 cloud-translation hours per month.
The publisher describes Pro's models as more accurate, but that is not a measured result for your recording. Evaluate a representative file, inspect the output, and check current model and device requirements before depending on a long transcription queue.