Get started

Tutorial · Python · ~10 min

How to Extract Hardcoded Subtitles from Video in Python

No CUDA setup. No GPU. No model weights. Just an API call that turns burned-in subtitles into clean, timestamped SRT files.

PythonOCRSRT
Extract subtitles

Why an API beats rolling your own OCR

Open-source OCR on video means CUDA configs, tile sizes, VRAM limits and accuracy tuning — then batch code on top. The API removes the GPU problem entirely. See the full comparison on the open-source tools page.

Step 1 — Get your API key

  • Sign up (free tier included)
  • Copy the key from the dashboard
  • pip install mlslabs

Step 2 — Use our sample video (or your own)

Python
import mlslabs

client = mlslabs.Client("YOUR_API_KEY")

Step 3 — Submit the extraction job

Python
job = client.subtitle_extractor.submit(
    input_url="s3://bucket/short-drama-cn.mp4",
    output_url="s3://bucket/subs/",
    languages=["zh", "en"],
    output_format="srt",   # srt | vtt | json
    confidence_threshold=0.6,
)
print(job.job_id)

Step 4 — Poll for the result

Python
job.wait()
print(job.status)
print(job.files)  # SRT + JSON sidecars

Step 5 — Inspect the SRT output

JSON output adds per-line bounding boxes, confidence scores and style attributes for downstream tools.

Text
1
00:00:02,120 --> 00:00:04,860
The door creaked open as she stepped inside.

2
00:00:04,860 --> 00:00:07,400
She stepped in and saw the letter on the desk.

Step 6 — Batch mode

For a whole episode library, submit one batch: queue thousands of files with per-file webhooks and automatic retries.

Run this tutorial in Postman — extraction request prefilled with the sample episode.

Keep exploring

Related guides

Subtitle Extractor API

Extract on-screen text and spoken dialogue with frame-accurate timestamps.

Learn more

Subtitle Eraser API

Remove hardcoded subtitles cleanly for localization and short-drama export.

Learn more

Vocal Separator API

Isolate vocals or instrumentals from speech and music mixes.

Learn more