Turkish subtitle generator
Turkish is Shycut’s second language, and its speech model gets 15.9% of Turkish words wrong with Accurate in OpenAI’s figures. Write the subtitles without uploading anything.
Drop your video here
Choose your videoMP4, MOV, WebM or MKV. It opens right here, on your device.
- Nothing uploaded
- No sign-up
- No watermark
How often it gets a Turkish word wrong
| Model in Shycut | Word error rate | Roughly | From the paper |
|---|---|---|---|
| Accurate (Whisper small) | 15.9% | 1 word in 6 | Table 13, “Whisper small” row, Turkish column · OpenAI |
| Fast (Whisper base) | 27.5% | 1 word in 4 | Table 13, “Whisper base” row, Turkish column · OpenAI |
Word error rate is the share of words a model gets wrong (swapped, missed or added) against a human transcript, in OpenAI’s own test on the FLEURS speech set; lower is better.
Shycut runs these two models converted to work in a browser (with 8-bit or 4-bit weights on some devices), so its results can differ a little from the paper’s.
Which model to pick
- Accurate
- Makes 42% fewer mistakes in Turkish than Fast. A one-time download of about 300 MB, kept on your device.
- Fast
- About 100 MB to download and quicker to run, at about 1 word in 4 wrong instead of 1 in 6.
- Either way
- Correct any line in the Captions tab before saving; the speech never leaves your device.
Turkish, with its own hesitations
The Fast model gets 27.5% of Turkish words wrong, about 1 in 4; Accurate cuts that by 42% to about 1 in 6. Long words built from many suffixes are where a wrong letter shows, so read them through.
Shycut knows Turkish hesitation sounds such as “ııı” and “eee”, and can cut them from the video along with silences. The subtitles and the cuts come from the same written speech.
Need English captions for Turkish speech? Translate it instead.
How to do it in Shycut
Open your video
The button opens Shycut with the captions set to Turkish, ready to write.
Write the captions
The speech model downloads once, then turns the speech into timed lines in this browser tab.
Check, then save
Fix names and terms, pick a caption style, and save the video with the captions in it, or download them as an SRT file.
Where these facts come from
Questions
How accurate is Turkish speech recognition?
In OpenAI’s test, 15.9% of words wrong for the Accurate model and 27.5% for Fast.
Can it cut “ııı” and “eee” from my video?
Yes. Shycut finds Turkish hesitation sounds and cuts them with one switch in the Cut tab, or as the video opens with “Cut “um”s and “uh”s” on the home page.
Can I get English captions for Turkish speech?
Yes, with the Accurate model; the Turkish translation page shows how well it translates.
Your video stays with you.
Open Shycut, drop the video in and export. Free, with no account, and your video never leaves your device.
Choose your video