Antimatter

GET/api/studio/modelsNo key

Every live video and sound job model with its options and its billed price at the defaults and at its dearest, markup on; whether chat's speech and picture doors have a priced model; and the media job markup. on is false while jobs are switched off.

Answer 200

The models.

Fields of the 200 answer
Field	Type	What it is
on	boolean
video	array of object
video[].id	string
video[].name	string
video[].maker	string
video[].sample	object	the clip the model made from a stated prompt, where the house made one
video[].sample.src	string
video[].sample.mediaType	string
video[].kind	string	One of "video", "sound", "still".
video[].input	string	One of "text", "image".
video[].references	object	Reference pictures it takes, the most any of its providers takes: how many at least and at most, and the one aspect ratio and length, or a resolution, it holds a request with any to. Absent: it takes none.
video[].references.min	integer
video[].references.max	integer
video[].references.aspectRatio	string
video[].references.seconds	number
video[].references.notAt	string
video[].endFrame	boolean	Whether it takes a last frame beside its first.
video[].trajectory	boolean	true where its card takes a drawn camera path as its own input; otherwise a shot's drawn path is sent as words in the prompt.
video[].seconds	object
video[].resolutions	object
video[].aspectRatios	object
video[].audio	object
video[].audio.default	boolean
video[].seed	boolean	true where the model takes a seed, so takes of one shot can differ by it alone
video[].unit	string	One of "second", "output".
video[].defaultMicros	integer
video[].worstMicros	integer
video[].perSecondMicros	integer
video[].holdersOnlyUntil	integer
sound	array of object
sound[].id	string
sound[].name	string
sound[].maker	string
sound[].sample	object	the clip the model made from a stated prompt, where the house made one
sound[].sample.src	string
sound[].sample.mediaType	string
sound[].kind	string	One of "video", "sound", "still".
sound[].input	string	One of "text", "image".
sound[].references	object	Reference pictures it takes, the most any of its providers takes: how many at least and at most, and the one aspect ratio and length, or a resolution, it holds a request with any to. Absent: it takes none.
sound[].references.min	integer
sound[].references.max	integer
sound[].references.aspectRatio	string
sound[].references.seconds	number
sound[].references.notAt	string
sound[].endFrame	boolean	Whether it takes a last frame beside its first.
sound[].trajectory	boolean	true where its card takes a drawn camera path as its own input; otherwise a shot's drawn path is sent as words in the prompt.
sound[].seconds	object
sound[].resolutions	object
sound[].aspectRatios	object
sound[].audio	object
sound[].audio.default	boolean
sound[].seed	boolean	true where the model takes a seed, so takes of one shot can differ by it alone
sound[].unit	string	One of "second", "output".
sound[].defaultMicros	integer
sound[].worstMicros	integer
sound[].perSecondMicros	integer
sound[].holdersOnlyUntil	integer
still	array of object	The models that make a shot's still, from words or reference pictures.
still[].id	string
still[].name	string
still[].maker	string
still[].sample	object	the clip the model made from a stated prompt, where the house made one
still[].sample.src	string
still[].sample.mediaType	string
still[].kind	string	One of "video", "sound", "still".
still[].input	string	One of "text", "image".
still[].references	object	Reference pictures it takes, the most any of its providers takes: how many at least and at most, and the one aspect ratio and length, or a resolution, it holds a request with any to. Absent: it takes none.
still[].references.min	integer
still[].references.max	integer
still[].references.aspectRatio	string
still[].references.seconds	number
still[].references.notAt	string
still[].endFrame	boolean	Whether it takes a last frame beside its first.
still[].trajectory	boolean	true where its card takes a drawn camera path as its own input; otherwise a shot's drawn path is sent as words in the prompt.
still[].seconds	object
still[].resolutions	object
still[].aspectRatios	object
still[].audio	object
still[].audio.default	boolean
still[].seed	boolean	true where the model takes a seed, so takes of one shot can differ by it alone
still[].unit	string	One of "second", "output".
still[].defaultMicros	integer
still[].worstMicros	integer
still[].perSecondMicros	integer
still[].holdersOnlyUntil	integer
speech	boolean
images	boolean
markupBps	integer
maxActiveJobs	integer	How many shots of an account's render plans run at once; the rest wait for a slot.
freeShot	object	Present while the first shot on us is on: the model a free shot is made with and its length in seconds. Whether the signed in account may take it is said by the quote (free_shot).
freeShot.model	string
freeShot.seconds	number
samples	object	The studio's own sample takes where the house made them: versioned urls under /api/studio/samples, by sample still id, by camera move id, and by the presets library's template id (an effect or a whole shot recipe).
samples.stills	object
samples.presets	object
samples.templates	object
feed	boolean	Present and true while the public feed of shared takes is on (/api/studio/feed).
storage	boolean	Whether stored media is on: the account keeps the files of its takes, characters and shares, so another device can bring a take in.
tasks	array of object	The task models, cheapest unit first: speech, lip sync, a talking still, edit, extend, upscale, motion control, and a picture's upscale and fill. One that starts from a file reads it from stored media, so it is listed only while stored media is on.
tasks[].id	string
tasks[].name	string
tasks[].maker	string
tasks[].task	string	One of "speech", "lipsync", "talking", "edit", "extend", "upscale", "motion", "enlarge", "inpaint".
tasks[].kind	string	What it makes: speech a sound, a picture's upscale (enlarge) or fill (inpaint) a still, every other task a video. One of "video", "sound", "still".
tasks[].unit	string	How the price is counted: per second of output, per job, per character read, or per megapixel (of the output for an upscale, of the still for a fill). One of "second", "output", "character", "megapixel".
tasks[].unitMicros	integer	Billed micro USD of one unit at the defaults: a second of output, the whole job, or a thousand characters. A task's total waits on the files it starts from, so the quote states it.
tasks[].worstMicros	integer	Billed micro USD of the dearest job it takes, at the longest files.
tasks[].adds	number	An extension only: the seconds each step adds after the clip it reads, as its card states. A clip grows in steps while it stays inside takes.video.maxSeconds.
tasks[].takes	object	The files it reads (video_asset, audio_asset, image_asset, mask_asset) with their limits, and whether it reads words.
tasks[].takes.video	object
tasks[].takes.video.minSeconds	number
tasks[].takes.video.maxSeconds	number
tasks[].takes.video.types	array of string
tasks[].takes.video.maxBytes	integer
tasks[].takes.video.minSide	integer
tasks[].takes.video.maxSide	integer
tasks[].takes.audio	object
tasks[].takes.audio.minSeconds	number
tasks[].takes.audio.maxSeconds	number
tasks[].takes.audio.types	array of string