List Percify API models and their inputs

GET https://api.percify.io/v3/playground/v1/models returns every model you can run through the Percify API, with the id to send to /v1/run, what it outputs, its credit cost and the JSON Schema of its inputs. GET /v1/models/{id} returns one model, plus a worked example when one is available. Neither call needs an API key. On 16 September 2026 the catalog listed 130 models: 72 video, 41 image and 17 audio. The list changes as models are added, so read it from the API rather than hard-coding it.

List all models

curl -s https://api.percify.io/v3/playground/v1/models

Get one model

string

required

A model id from the list, for example infinitetalk-fast.

curl -s https://api.percify.io/v3/playground/v1/models/zonos2

An unknown id returns 404 with Model not found. Available models: ….

Model fields

string

The value to send as modelId.

string

Display name, for example Zonos 2.

string

What the model does.

string

image, video or audio.

string

The catalog group the model is listed under.

integer

The model’s listed credit cost. When pricing is null, this is the price of every run.

object | null

Present when the price depends on the input. type is the billing method and varies_by names the inputs that change the price. See the table below.

string | null

A preview image URL.

object

JSON Schema for input: required fields, and for each property its type, enum options, default and a description.

object

Only on GET /v1/models/{id}, and only when an example exists: output.type and output.url of a real result, and sometimes the input that made it.

object

The run and poll URLs for this model.

Example (trimmed)

{
  "success": true,
  "data": {
    "id": "zonos2",
    "name": "Zonos 2",
    "output_type": "audio",
    "category": "audio",
    "credits": 2,
    "pricing": { "type": "per_period", "varies_by": ["_audioDurationSec"] },
    "input_schema": {
      "type": "object",
      "required": ["text", "audio"],
      "properties": {
        "text": { "type": "string", "description": "What the cloned voice should say" },
        "audio": { "type": "string", "format": "uri", "description": "Short.mp3/.wav sample of the voice to clone (public https URL)" },
        "clean_speaker_background": { "type": "boolean", "default": false }
      }
    },
    "endpoints": {
      "run": "POST https://api.percify.io/v3/playground/v1/run",
      "poll": "GET https://api.percify.io/v3/playground/v1/generations/{id}"
    }
  }
}

Pricing types

`pricing.type`	How the run is priced
`null` (no pricing object)	A flat `credits` per run.
`per_second`	By the seconds of an input: the audio length (`_audioDurationSec`) for lip-sync, or a `duration` option for video. Some also multiply by `resolution`.
`per_period`	By started periods of a length, such as the length of a voice sample, or scaled by an option such as `quality`.
`per_character`	By the length of the text input.
`matrix`	A price for each combination of two options, such as quality and size.

_audioDurationSec is not an input you send. Percify measures the audio or video file you pass. For any model with a pricing object, POST /v1/estimate gives the exact price.

Models for common jobs

Job	Model ids
Talking avatar video from a photo and audio	`infinitetalk-fast`, `infinitetalk`
Clone a voice and speak a script	`zonos2`
Text to speech with a preset voice	`speech-02-hd`, `speech-02-turbo`, `chatterbox-turbo`
Images from a prompt	`gpt-image-2`, `flux-schnell`, `nano-banana-pro`

These ids were in the catalog on 16 September 2026. Check GET /v1/models for the current list.

Other machine-readable formats

OpenAPI 3.1: api.percify.io/v3/playground/v1/openapi.json describes /v1/run and /v1/generations/{id}, with each model’s input schema under components.schemas.Input_<modelId>.
Plain text for agents: api.percify.io/v3/playground/v1/llms.txt lists each model id, its output type and required inputs.

Start a generation

Send a model id and its inputs.

Estimate cost

Exact credits for input-priced models.

Talking avatar video

Chain an image, a voice and lip-sync.

Voice and speech

Voice cloning and text to speech models.

Get started

Generations

Async and webhooks

Build with the API

Account

MCP

List Percify API models and their inputs

List all models

Get one model

Model fields

Pricing types

Models for common jobs

Other machine-readable formats

Start a generation

Estimate cost

Talking avatar video

Voice and speech

​List all models

​Get one model

​Model fields

​Pricing types

​Models for common jobs

​Other machine-readable formats

​Related

Start a generation

Estimate cost

Talking avatar video

Voice and speech

List all models

Get one model

Model fields

Pricing types

Models for common jobs

Other machine-readable formats

Related