How it works
Point it, run it, review it.
openvideo does not ask you to reorganise anything. Your videos stay where they are: openvideo processes them there and writes the text next to them.
- 1
Point it at your videos.
In the app you register the network storage and the folder where your recordings are. You can browse folders and files, and see which videos already have text.
- A folder with all its subfolders, or only the files you select.
- At a glance you see what is done and what is not.
- A single new recording can also be uploaded from the browser into a source folder.
- 2
Run it.
Choose which files to produce (all of them by default) and set the language of the recordings. To be on the safe side, try the settings on one video first, review it, and only then start the rest.
- Live progress: how many files and how many hours are done, and how much is left.
- Pause, resume, stop, and retry the files that failed.
- Later, as the system improves, you can run it again at any time.
- 3
Review.
Every video has its own page: the transcript, the summary, the corrections and the uncertain words, and the files to download. Once a person has read it and found it right, they mark it as reviewed.
- You see what the system corrected and where it was unsure.
- A reviewed transcript is never overwritten by a later run.
- The files sit next to the video, so your usual tools open them too.
What it writes next to the video
Every file takes the video’s own name, so you always know what belongs to what.
Readable transcript
The full text with speaker changes and timestamps, opening with the AI summary, the keywords and the chapters.
Subtitles
In two formats (SRT and SAMI), named after the video. The usual players load them by themselves when they sit next to the video.
Plain text
Just the words, no formatting: for search, quoting and other systems.
Structured file
Every word, timestamp and mark, for machine processing: when another system works on with the text.
Accuracy, honestly.
openvideo writes a verbatim transcript: filler words stay in, and nothing is paraphrased. Where it is unsure of a word, it marks the word instead of writing a plausible-sounding one in its place. Every correction can be traced. What the final text is, a person decides. The marks never make it into the subtitles.
Hungarian first, English too.
openvideo is built for Hungarian speech first, and English recordings get an English transcript. You set the language for the recordings; the system never guesses it.
See it at work.
A personal online demo, with your questions, on the kind of recordings you work with.