Describe Image
Midjourney image-to-text (DESCRIBE) endpoint
Uploads an image and asks Midjourney to reverse-engineer (DESCRIBE) its prompt. The submission is async: on success it returns a task ID (result), and you read the generated description via Query Task Result.
Billing model
This endpoint is billed under the mj_describe model. Configure a group price for it on the channel, and your token must be allowed to use that model's group.
How to send the image
Use either form: prefer base64Array with a Data URL (data:image/...;base64,...), or content with a publicly reachable http(s) image URL.
Authorization
BearerAuth
使用你在 OmwAI 中创建的 API 令牌认证。
格式: Authorization: Bearer sk-xxxxxx
令牌需开通对应 Midjourney 模型的调用权限。
In: header
Request Body
application/json
待识别图片的 Data URL 列表（取第一个元素），与 content 二选一
待识别图片的公网 http(s) URL，与 base64Array 二选一
自定义状态字段，查询结果时原样返回
任务完成后的回调通知地址（若系统启用回调通知）
Response Body
application/json
curl -X POST "https://www.omwai.xyz/mj/submit/describe" \ -H "Content-Type: application/json" \ -d '{}'{
"code": 1,
"description": "success",
"result": "31f07c2a-1d20-4a53-9b84-3c1b2d4e5f60",
"properties": {}
}Note
A successful submission (code of 1) does not mean the result is ready. Poll the Query Task Result endpoint with the returned result (jobId) until status becomes SUCCESS; the description field then contains the reverse-engineered prompt(s).
How is this guide?
Last updated on