Extract Subtitles for Translation

Why this matters

Machine translation works on text. When the only copy of dialogue is burned into pixels, every language track starts with manual transcription.

Frame-OCR plus temporal clustering produces timed, deduplicated SRT from video alone — no source script needed.

How it works

The API OCRs frames, clusters identical text across time, collapses duplicates, and emits SRT/VTT with confidence scores per line.

Submit a job with an input URL (S3, GCS or HTTPS), poll the job URL, and download the rendered output. No GPU, no queues, no ffmpeg builds to babysit.

Code

import requests, time

API = "https://api.mlslabs.io/v1/subtitle-extractor/jobs"
headers = {"X-API-Key": "YOUR_API_KEY"}
payload = {"input": "s3://bucket/episode.mp4", "output_format": "srt", "lang": "zh"}

resp = requests.post(API, json=payload, headers=headers)
job = resp.json()
while job["status"] not in ("succeeded", "failed"):
    time.sleep(3)
    job = requests.get(job["url"], headers=headers).json()

print("Output:", job["output_url"])

Pro tips

• Prefer the cleanest episode master for OCR accuracy.
• Hand the confidence scores to your QC step, not the translators.
• For zh → en MT, feed extracted SRT directly; no re-timing needed.

Run the Subtitle Extractor collection — episode sample prefilled.