Captions, ReCreate, HDR, Avatar and Influencer by API

Make captions, ReCreate pictures, HDR photos, avatar videos and AI Influencer videos through the API, and what each one needs first.

Open the tool

This page covers the tools of the Public API that take more than one step on our side.

What it does

Each call answers at once with a job id, like any other job. ReCreate, HDR, Avatar and Influencer jobs start within a few seconds of your call, so their status is queued first. If one cannot start, it ends failed, its error says why, and nothing is charged.

GET /api/public/v1/me lists in tools what this key can run now. Draw your screen from that list.

Every option explained

Captions

POST /api/public/v1/captions

Field	Meaning
`asset`	The asset id of a clip or a sound file
`language`	A language code. Leave it out to detect it.

A finished job has files: srt, and words with every line and its timing. The price is per minute of sound.

Caption overlay

A clear video of your captions in a Kriti Caption style, to lay over an edit. It costs no credits, and it needs a plan with the transparent export.

GET /api/public/v1/caption-styles lists key, name, chip, pro and locked.

Field	Meaning
`style`	A style key. `chamak` by default.
`width`, `height`	The frame size, in even numbers
`fps`	The frame rate
`format`	`prores` by default, or `webm`
`from_ms`, `to_ms`	Only this part of the sound, for a trimmed clip. Times start from `from_ms`.

One overlay runs at a time for each account, with a daily count and a length limit, as on the website. The file stays for a day.

ReCreate

POST /api/public/v1/recreate makes one picture from a reference.

Field	Meaning
`reference`	The asset id of the picture to follow
`rights`	Must be `true`: you confirm you may use this reference
`ratio`	`1:1` by default
`tier`	`studio`, `pro` or `pro_hd`
`changes`	A list of what to change, in your own words

To swap in your own model or animate the result, use the website.

HDR Processing

POST /api/public/v1/hdr takes asset, the asset id of one photo, and gives back the finished picture.

AI Avatar

GET /api/public/v1/avatars lists the presenters you can use. Then POST /api/public/v1/avatar:

Field	Meaning
`presenter`	A presenter id
`presenter_kind`	`avatar`, `me` or `influencer`, from the same list
`script`	What to say
`language`	`hi` by default
`voice`	A voice id. Leave it out for the presenter's own voice.
`ratio`	`1:1` by default

Your own face needs the consent tick on the website first.

AI Influencer

GET /api/public/v1/influencers lists the ones ready to make a video. Then POST /api/public/v1/influencer:

Field	Meaning
`influencer`	Its id
`script`	What to say, up to 2,000 characters
`look`	An approved Look. Leave it out for the default one.
`language`	`hi`, `en` or `hg`

Making a new influencer stays on the website.

Common problems

I get overlay_busy

An overlay is still running. Wait for it to finish.

I get influencer_not_ready

Lock its identity and approve a Look on the website first.

I get consent_required

Open AI Avatar on the website once with this face and tick the consent.

My job says queued for a while

These jobs wait for a worker. It is usually seconds, and a minute at most when many are waiting.

Related guides

Updated 3 Nov 2026