01
Pixel-level burn-in OCR
No soft track? Fine. We sample frames, read the caption band, and reconstruct timed lines from the image itself.
Hardcoded only
Capabilities
Everything between burn-in and a clean file
subtitleOCR is a single job: read hardcoded captions, rebuild timed lines, export SRT or TXT—across 75+ languages.
02
75+ language packs
Pick a preset—or the bilingual English + Simplified Chinese mode when both scripts share a frame.
Multilingual
03
Region + range control
Draw the band, set language, and limit the process window so you scan only the subtitled stretch you care about.
Pay for minutes used
04
Timeline before commit
Every line lands with start, end, and a plain-text preview. Spot-check before you download.
Review first
05
SRT & TXT exports
Copy or download either format. Guests get a free preview cut; accounts unlock the full file plus 2 free hours on signup.
Editor-ready
06
Local library + account history
Browser history keeps full results handy. Signed-in jobs also show credit usage and status on the dashboard.
Come back later
Accuracy is a lever, not a lottery
Contrast, typeface, and band placement decide the quality of a read. subtitleOCR gives you the controls—language pack, custom region, process range—so messy or mixed frames get a second chance instead of a shrug.
- →Tighten the region when text sits off-center
- →Switch language packs for non-Latin scripts
- →Crop the process range to skip empty lead-ins
Clip on disk? Unburn it.
Open the studio and pull captions out of the picture.