FFmpeg drawtext on a hosted API: why Sume filters refuse it, and cues

Sume's video filter does not accept drawtext: it is not on the allowlist, and the docs say drawtext, subtitles, movie and lut3d are out, along with anything that reads a file or socket. To put your own text on a clip, send cues to video captions: each cue is text, start and end in seconds, and Sume burns exactly that copy at those times with no speech-to-text.

FFmpeg's description of drawtext is from its filters documentation, read 2026-10-02.

Why is drawtext refused?

FFmpeg's entry says drawtext draws a text string, or text from a specified file, using libfreetype, and takes a font setup. The Sume filter list is limited to filters with only inline options so that nothing on it can read a path. An off-list name returns invalid_filtergraph naming the allowed filters. A free POST /v1/video-filter/check shows that before you pay.

How do cues work?

cues (alias segments) holds 1 to 200 items; each has text up to 400 characters, and start and end from 0 to 60 seconds. Text may contain a newline for a two-line card. cues cannot be combined with words, segments or script_text. Pick a style (slam, punch, tiktok-green, or a Hangul style) and tune it with design. A standalone caption job is $0.20 for videos up to 60 seconds.

The video_url must be a fetchable public HTTPS URL. Poll GET /v1/jobs/:id/status and read the result, or GET /v1/video-captions/:id.

curl -X POST https://api.sume.com/v1/video-captions \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: cue-text-001" \
  -d '{
    "video_url": "https://media.sume.com/artifacts/example/clean.mp4",
    "style": "punch",
    "cues": [{ "text": "Free shipping today", "start": 0.5, "end": 3 }]
  }'

When is cues the wrong tool?

A silent clip with no cues fails as caption_no_speech; the fix is the cue form above.