Skip to content
Log in
Developer API
Get your keyGet a key

Reference · Video

POST/videos/quote

Price it: make a clip from words, frames, references or a video to follow

The price for exactly this body, the balance beside it and any notes. Nothing runs and nothing is charged. Run it with POST /api/v1/videos, the same body, confirmCredits set to credits and the same Idempotency-Key.

Key needs
create · Make things and spend credits
Runs it
POST /videos
Answers
A quote · free
Request
curl -X POST https://vessero.com/api/v1/videos/quote \
  -H "Authorization: Bearer $VESSERO_API_KEY" \
  -H "Idempotency-Key: 47f8f6f5-ea00-4fb1-925f-76aac956a0ec" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedance-2.5",
    "prompt": "A slow push in on a steaming coffee cup on a café table",
    "durationSec": 5,
    "firstKey": "https://example.com/frames/coffee-cup.jpg"
  }'
Response200 OK
{
  "object": "quote",
  "credits": 12,
  "balance": 840,
  "lines": [
    "1 image × 12 (Nano Banana Pro · 4K)"
  ],
  "notes": [],
  "requestId": "47f8f6f5-ea00-4fb1-925f-76aac956a0ec",
  "run": {
    "method": "POST",
    "path": "/api/v1/images",
    "confirmCredits": 12,
    "idempotencyKey": "47f8f6f5-ea00-4fb1-925f-76aac956a0ec"
  }
}

Header

  • Idempotency-KeystringYour id for this request: 8–64 letters, digits, - or _. A repeat within 10 minutes gets the first answer back and never runs twice. Use the same one for the quote and the run.

Body

  • promptstringWhat happens in the shot, e.g. "Slow push in on the bottle as morning light sweeps across the table". For an edit, the change to make; to extend, what happens next. Optional only for motion, whose moves come from the video.
  • modelstringrequiredA video model id from GET /api/v1/models (studio "video"), e.g. "seedance-2.5".
  • durationSecintegerSeconds, one of the model’s durations. Default: its own length, as /video starts it (defaultDurationSec in GET /api/v1/models).
  • aspectRatiostringA shape the model lists, e.g. "9:16". Default: the model’s own.
  • resolutionstringA size the model lists, e.g. "720p". Default: the model’s own.
  • audiobooleanSound, on a model where it is optional (sound "optional"); it may cost more.
  • musicbooleanBackground music of the model’s own, on a model with a Music switch (music "optional" in GET /api/v1/models: Seedance), while sound is on. false: no background music — only the voices and sounds the prompt asks for. true: the model may add its own. Default: false, as the Video page starts it. Never moves the price.
  • draftbooleanA draft, on a model with the switch (draft in GET /api/v1/models: Seedance 2.5): made at its draft resolution (480p) and that price, a quick look. The clip worth keeping is made sharp with upgradeDraftOf, within its days. Default: off.
  • loopbooleanA clip that loops, on a model that makes one (loop "optional" in GET /api/v1/models: Grok Imagine 1.5): the start frame is the end frame too. Only with firstKey and no lastKey. Never moves the price.
  • cutsbooleanCuts, on a model with the switch (cuts "optional" in GET /api/v1/models: PixVerse v6, Kling O3 (reference)): the model plans a few shots in one clip — never between firstKey and lastKey, beside shots of the person’s own, or with a reference clip. Never moves the price.
  • seedintegerA seed, on a model that takes one (seed "optional" in GET /api/v1/models: PixVerse v6, v6 Max): the same seed and settings help make the same clip again. Default: a new one each run. Never moves the price.
  • effectIdstringAn effect, by its id from GET /api/v1/models (effects, on PixVerse v6): a ready-made motion — a product turned into a landmark, a dance trend, a stadium moment — from the person’s own pictures in effectKeys. The effect fixes the clip’s length, and its price follows it; send no durationSec, frames or references with it. Words are optional: without them the effect’s name goes.
  • effectKeysstring[] · up to 2The effect’s pictures, as many as it takes (its pictures in GET /api/v1/models: 1 or 2), in order.
  • referenceLinkstring · httpsA public web page the clip is made from, by its https:// link, on a model that reads one (page "optional" in GET /api/v1/models: Wan 3.0, Wan 3.0 Prime): one that opens without signing in, read by the model for what the clip shows. Never with firstKey or lastKey. Never moves the price.
  • upgradeDraftOfstringA finished draft of theirs, by its id (from this tool or GET /api/v1/work, which says which clips are drafts): make it sharp — the same take at its model’s upgrade resolution (1080p), priced at that rate for its length — rather than a new clip. Send it with the draft’s model alone.
  • modelModestringThe way the model runs, on one that runs more than one way (ways in GET /api/v1/models): Kling 4.0’s keyframes — firstKey, keyframeKeys and lastKey, in that order — or omni, its references; Grok Imagine 1.5’s frames (its start, end and frames between, or a loop) or refs. Default: the way what is sent calls for: references run on the way that takes them.
  • firstKeystringThe start frame, a picture: a storage key from POST /api/v1/files or POST /api/v1/library/use, or an https link to the file (brought into their Uploads first).
  • lastKeystringThe end frame, a picture, on a model that takes one: a storage key from POST /api/v1/files or POST /api/v1/library/use, or an https link to the file (brought into their Uploads first).
  • keyframeKeysstring[] · up to 8Frames between firstKey and lastKey, in order, on a way that takes them: on Kling 4.0’s keyframes up to 8, and no more frames in all than one more than the clip’s seconds; on Grok Imagine 1.5’s frames up to 4, one fewer than its seconds, each at an even whole second. Named @Keyframe1… in the words.
  • referenceKeysstring[] · up to 30Reference pictures, up to the model’s own number (inputs.pictures).
  • referenceVideoKeysstring[] · up to 10Reference clips, up to the model’s own number (inputs.videos). Priced by their length where the model bills it.
  • referenceAudioKeysstring[] · up to 10Reference sounds, up to the model’s own number (inputs.sounds; its audioRefs says what each takes). On a model that takes any sound (Seedance, Wan 3.0) each is what the prompt makes it, named @Audio1, @Audio2… in the order sent: a voice whose timbre, tone and accent the clip speaks in, a song or a bed under it, or a sound effect. Never lip-synced to and never transcribed — write what is said in quotes in the prompt. For a recording said as it is, lip-synced, use POST /api/v1/avatars.
  • whoseAudioobject · kind no-voice | voiceWhose voice is in the reference sounds, as the person says: {kind: "no-voice"}, or {kind: "voice", name, agreed} for a real person’s. Not asked of sounds made in Vessero.
  • parentIdstringA finished clip of theirs to work on, by its id (from this tool or GET /api/v1/work). Give this or sourceVideoKey.
  • sourceVideoKeystringA video to work on, uploaded: a storage key from POST /api/v1/files or POST /api/v1/library/use, or an https link to the file (brought into their Uploads first).
  • sourceStartSecnumberWhere in the video to start, in seconds.
  • sourceEndSecnumberWhere in the video to end, in seconds.
  • clipUsestringWhat to do with the video put in (parentId or sourceVideoKey), on a model that does it (worksOnVideo): extend it, edit it (the default), or motion — the character on firstKey takes its moves.extendeditmotion
  • keepSoundbooleanAn edit: keep the video’s own sound (the default, where the model can) or make new.
  • motionFacingstringMotion: the character faces as in the clip or as in the picture.clippicture
  • motionBackgroundstringMotion: the clip’s background or the picture’s.clippicture
  • whoIsInPictureobject · kind no-person | personMotion: who is in the picture on firstKey, as the person says; asked of an uploaded picture.
  • whoIsInVideoobject · kind no-person | personWho is in the uploaded video, where the model keeps its people (an edit that keeps them).
  • shotsobject[] · up to 6A clip in shots, on a model that takes a storyboard (inputs.storyboard): its own number of shots, each its own seconds. On Kling O3 (reference) the shots go beside reference pictures (referenceKeys), and no clip goes with them.
    • promptstringrequiredWhat happens in this shot.
    • durationSecintegerrequiredIts seconds.
  • presetIdstringThe style a job runs in, on a model that has styles (presets in GET /api/v1/models: Restyle): one of its presets.choices ids. Needed where presets.required.
  • camerastringA camera move added to the words: dolly-in (Dolly in), dolly-out (Dolly out), crash-zoom (Crash zoom), orbit-360 (360 orbit), crane-up (Crane up), whip-pan (Whip pan), fpv-dive (FPV dive), handheld (Handheld), bullet-time (Bullet time), snorricam (Snorricam), drone-reveal (Drone reveal), locked-off (Locked-off). Never changes the price.dolly-indolly-outcrash-zoomorbit-360crane-upwhip-panfpv-divehandheldbullet-timesnorricamdrone-reveallocked-off

Answers

  • 200quoteThe price for exactly this body, your balance and any notes. Nothing ran.
  • 400invalid_requestA field is missing or not accepted: see details.
  • 403feature_unavailableThe studio this needs is switched off for the account right now.
  • 429rate_limitedToo many calls this minute: see Retry-After.
Next in VideoPOST/avatars