Canadian Thanksgiving hours clip in English and French for $0.40

To make a Canadian Thanksgiving hours video in English and French, render one clean clip with no text on it, then send two caption jobs to Sume's video captions endpoint, one with English cues and one with French cues. Each accepted caption job costs $0.20 for a clip of 60 seconds or less, so both language versions cost $0.40 on top of the clip itself. Canadian Thanksgiving falls on Monday, October 12, 2026.

Why does a Thanksgiving hours clip need its own cue set?

Because the hours are the whole message, and they are not the same everywhere. CTC News reports that hours can vary by province, municipality and individual organization, that Ontario applies retail-closing rules with exemptions, and that Nova Scotia treats the day as a non-statutory holiday for paid-holiday purposes. A shop with two locations in different provinces may therefore need two different clips.

A clip with the hours burned in as text also works with the sound off. Sume's caption endpoint can burn authored text without any speech, so a silent product clip is a valid input.

How do cues work without speech?

Speech-to-captions only works when the clip has audible speech; a silent clip fails with caption_no_speech. To burn your own words, send cues (or segments), each with text, start and end in seconds. Sume does not run speech-to-text on that path and burns the text you wrote at the times you gave. You can send only one of script_text, words, cues and segments per job.

Here is the English request. The French job is identical except for the three text values and the idempotency key.

curl -X POST https://api.sume.com/v1/video-captions \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: thanksgiving-hours-2026-en" \
  -d '{
    "video_url": "https://media.sume.com/artifacts/example/clean.mp4",
    "style": "slam",
    "cues": [
      {"text": "Open Monday, Oct 12", "start": 0.5, "end": 2.5},
      {"text": "10 am to 4 pm", "start": 2.5, "end": 4.5},
      {"text": "Happy Thanksgiving", "start": 4.5, "end": 6}
    ]
  }'

What does the whole job cost?

If you shoot the clip on a phone, the only Sume cost is the two caption jobs. If you want a generated backdrop, the Wan 3.0 rate on Sume is per output second and depends on resolution. The table prices a 6-second clip from the rate card in the repository (Wan 3.0 at 480p, 720p and 1080p), then adds the two caption jobs.

How do you check the French text before it goes out?

The caption docs state a hard failure for Korean text on Latin-only styles, but they do not list accented Latin characters, so do not assume every accent renders in every style. Pull a still at the moment each French cue is on screen with Sume's video frames endpoint, which returns durable images at times you name, and look at the accents before you post. The frames job is billed by its Modal compute rather than a flat rate.

If the look needs to change, send source_caption_id instead of video_url. Sume reuses the source video of that caption, and the price stays at a caption render.

• Keep each cue short, a few words, so the hours stay readable on a phone.
• Use the same start and end times in both languages so the two clips stay in step.
• Give each language its own Idempotency-Key so a retry never starts a second render.