Z-Image

Z-Image models are called through one unified, asynchronous job API. Start a task with POST /api/generate, then poll GET /api/generate/{taskId} until it finishes. This family exposes 1 endpoint.

Overview

Every endpoint shares the same envelope — only model and the input fields change. Run any from the Models Market playground.

Authentication

Create an API key at /api-keys, then send it as a Bearer token. Requests from the browser while logged in also work via the session cookie.

Authorization: Bearer <YOUR_API_KEY>

Create Task

POST

POST https://api.lithovas.com/api/generate

Body: { "model": "<slug>", "input": { … } }

Response: { "taskId": "<id>", "state": "waiting" }

Query Result

GET

GET https://api.lithovas.com/api/generate/<taskId>

Response: { "taskId", "state": "waiting" | "success" | "fail", "resultUrls": [ … ], "creditsConsumed", "failMsg", "balance" }

Asynchronous

A 200 from Create Task only means the task was accepted — not finished. Poll this endpoint until state becomes success or fail. Failed tasks are not billed.

Endpoints

Model	Task	model slug
Z Image	Text to Image	z-image/z-image

Z Image · Text to Image

POSTText to Imagez-image/z-image

input fields

promptstringrequired

Text description of the image to generate (max 1000 characters)

aspect_ratioenum<string>required

Aspect ratio for the generated image

nsfw_checkerbooleanoptional

Content filtering toggle; set false to disable filtering

Example

cURL

curl -X POST https://api.lithovas.com/api/generate \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "z-image/z-image",
  "input": {
    "prompt": "A cinematic red panda surfing through neon rain at dusk.",
    "aspect_ratio": "1:1",
    "nsfw_checker": false
  }
}'