How do I turn my free course lessons into text for a sales page?

To turn free course lessons into text for a sales page, detach the audio from each lesson video with POST /v1/audio-detach, then send it to POST /v1/stt-1.0/transcribe and read text from the result. On Sume that is $0.01 for the detach and $0.01 per audio minute for the transcript, so an 8-minute lesson is $0.09 and ten free-preview lessons are $0.90. Rates were read on 2026-10-07.

The transcript is raw material, not the page. A sales page needs the three or four sentences that show what a student learns, and you choose those. What the transcript gives you is exact wording, so you quote your own lesson and do not paraphrase it from memory.

Why detach first

The transcript job takes audio, and a lesson is a video. Audio detach takes one video that is already on media.sume.com and returns a new audio file. The video does not change. If the lesson lives somewhere else, import it first with POST /v1/media-imports, because the server does not fetch from the open internet.

Ask for format: wav, channels: mono and sample_rate: 16000, which the docs call the speech-to-text shape. Each detach is one flat cent and needs an Idempotency-Key. The default mode is async, so you get a job id and poll GET /v1/jobs/:id/status, or pass mode: sync to wait up to 30 seconds.

curl -X POST https://api.sume.com/v1/audio-detach \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: lesson-03-detach" \
  -d '{
    "video_url": "https://media.sume.com/artifacts/artf_demo/lesson03.mp4",
    "format": "wav",
    "channels": "mono",
    "sample_rate": 16000
  }'

Transcribe, and mind the 10-minute cap

One transcript job takes at most 10 minutes (600 seconds) of audio. Pass duration_seconds rounded up to the whole minute, because if you leave it out Sume reserves one minute. For an 8-minute lesson send 480.

Lessons longer than 10 minutes need slicing. Detach the whole track (a single detach output can be up to 900 seconds), cut it with Timeline audio split at $0.01 into ranges of 600 seconds or less, and send one transcript job per slice. Add each slice's start time to its word times when you merge them. A 12-minute lesson then costs $0.01 to detach, $0.01 to split and $0.12 for 12 audio minutes: $0.14.

If you want to keep lessons under 10 minutes anyway, the free preview is easier to watch and cheaper to process.

curl -X POST https://api.sume.com/v1/stt-1.0/transcribe \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: lesson-03-stt" \
  -d '{
    "audio_url": "https://media.sume.com/artifacts/artf_demo/lesson03.wav",
    "language_code": "en",
    "duration_seconds": 480,
    "segmentation": "sentence"
  }'

What to do with the text

Read each transcript once against the video, and fix names, product terms and numbers by hand. Speech-to-text is a draft, and a sales page repeats it under your name.

Pull out three things per lesson: the problem the lesson solves, the one step it teaches, and the result the student can show. Put those under each lesson title on the page. Keep the sentence timings from segments[], so each quote can link to the right minute in the video.

The same transcript can also become captions. Pass it as script_text to POST /v1/video-captions ($0.20 per job) and the words are aligned to the speech, so the free lessons are accessible without a second transcript.

Keep the transcript honest

A sales page that quotes your lessons must quote them correctly. Read each transcript against the video once, fix the names of tools and people, and check every number you plan to repeat. If a lesson mentions a price, a date or a result that has changed since you recorded it, do not quote it, or add the current figure beside it.

Do not claim more than the lesson teaches. If the lesson shows one way to do a thing, say that, not that it teaches 'everything about' the topic. Students who find the page matches the lesson are the ones who ask for a refund least.

Store each transcript next to its lesson id with the date you made it. If you re-record a lesson, run the two calls again for about nine cents, and replace the page text the same day.

Costs for a free-preview module

• Rates apply per job and per audio minute. Check the live rate in GET /v1/catalog before a bulk run.
• A retry with the same Idempotency-Key does not create a second paid job.
• Keep the transcript with the lesson id, so you can regenerate the page when a lesson is re-recorded.