Create PDF Parsing Task (Official Free Version)

POST

/mineru/api/v4/extract/task

Document Reference: https://mineru.net/apiManage/docs

📌 Function Description
This API is used to parse documents corresponding to specified URLs. It supports functions such as OCR recognition, formula/table extraction, and multi-format export. Parsing results can be obtained through callback notifications or scheduled polling, making it suitable for structured parsing scenarios of various types of documents.

⚠️ Pre-Use Notes

For HTML file parsing, model_version must be explicitly specified as "MinerU-HTML". For non-HTML files, you can choose the pipeline or vlm version according to your needs.

The size of a single file cannot exceed 200MB, and the number of pages must not exceed 600.

Due to network restrictions, requests for foreign URLs such as those hosted on GitHub and AWS may time out.

This API does not support direct file upload.

📝 Usage Steps

Set basic parameters in this API and send a request to obtain the task_id.

Go to the Task Query Interface and fill in the task_id to get the results.

🔑 Key Precautions

Mandatory parameter verification: url and model_version are mandatory parameters; the lack of either will cause the API call to fail. When using callback, the seed parameter must be provided together.

Page range format: Supports single page numbers, continuous page ranges (e.g., 4-6), reverse page numbers (e.g., -2 indicates the second-to-last page), and multiple rules separated by commas.

Export format description: Markdown and JSON formats are exported by default. Additional export formats only support one or more of docx, html, and latex, which are specified through the extra_formats array.

💰 Price
Free for a limited time. The daily quota is limited and will be unavailable once used up.

Request

Header Params

Body Params application/json

Examples

vlm modeMinerU-HTML Mode (Parsing HTML Files)

{
    "url": "https://cdn-mineru.openxlab.org.cn/demo/example.pdf",
    "model_version": "vlm"
}

Request Code Samples

Shell

JavaScript

Java

Swift

PHP

Python

HTTP

Objective-C

Ruby

OCaml

Dart

cURL

vlm modeMinerU-HTML Mode (Parsing HTML Files)

curl --location --request POST 'https://api.302.ai/mineru/api/v4/extract/task' \
--header 'Authorization: Bearer ' \
--header 'Content-Type: application/json' \
--data-raw '{
    "url": "https://cdn-mineru.openxlab.org.cn/demo/example.pdf",
    "model_version": "vlm"
}'

Responses

🟢200success

application/json

Body

Example

{
    "code": 0,
    "msg": "ok",
    "trace_id": "79dd6fcac0f4b7e0a21fcc56854329a2",
    "data": {
        "task_id": "a4379d13-95e2-4af1-aae4-04fe5d6fc4cf"
    }
}

Modified at 2026-01-23 02:52:21

View PDF Parsing Task (Open Source Deployment Version)

View PDF Parsing Task (Official Free Version)