Skip to main content
Rendobar provides two caption-related job types: caption.extract for speech-to-text transcription, and caption.burn for overlaying subtitles onto video.

​
Extract captions (speech-to-text)

The caption.extract job uses AI-powered transcription (Deepgram) to convert spoken audio into timed SRT subtitles. 
curl -X POST https://api.rendobar.com/jobs \
  -H "Authorization: Bearer rb_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "caption.extract",
    "inputs": {
      "source": "https://example.com/video.mp4"
    },
    "params": {
      "language": "en"
    }
  }'

​
Parameters

ParameterTypeDefaultDescription
languagestring"en"Language code (ISO 639-1). Supports 30+ languages.
modelstring"nova-3"Deepgram model. Options: nova-3 (best quality), nova-2, base

​
Output

The result is an SRT file containing timed captions: 
1
00:00:01,000 --> 00:00:04,500
Welcome to this tutorial on video processing.

2
00:00:05,200 --> 00:00:08,800
Today we'll cover watermarking and captioning.
Retrieve the captions from the job result URL.

​
Burn captions onto video

The caption.burn job overlays subtitles directly onto the video frames, producing a new video with hard-coded captions. 
curl -X POST https://api.rendobar.com/jobs \
  -H "Authorization: Bearer rb_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "caption.burn",
    "inputs": {
      "source": "https://example.com/video.mp4",
      "subtitles": "https://example.com/captions.srt"
    }
  }'
If you omit the subtitles input, Rendobar will automatically extract captions first and then burn them — combining both steps in a single job. 
# Auto-extract + burn in one step
curl -X POST https://api.rendobar.com/jobs \
  -H "Authorization: Bearer rb_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "caption.burn",
    "inputs": {
      "source": "https://example.com/video.mp4"
    }
  }'

​
Subtitle format

Rendobar accepts SRT format. Each block contains a sequence number, timestamp range, and text: 
1
00:00:01,000 --> 00:00:04,500
Your caption text here.

​
Two-step workflow

For maximum control, use a two-step workflow:
1

Extract

Extract captions with caption.extract.
2

Review and edit

Download the SRT file and fix any transcription errors or adjust timing.
3

Burn

Submit the edited SRT file with caption.burn to overlay onto the video.
This gives you full control over caption accuracy before committing to the final video.

​
Cost

Job TypeCost Model
caption.extract$0.50/min of input
caption.burn$1.00/min of output

Job Types

Full list of operations

GDPR Guide

Redact faces alongside captions