Use the Document Parsing Console

In the document parsing console, you can select files, adjust parsing settings, and check text-and-image or audio/video results against the original files.

Enter the Parsing Workspace Open ParseX, log in, and select "Document Parsing" on the left. The workspace includes input, configuration, and result areas, and you can complete a full try-out on the same page.

Workspace section	Purpose
Input file	Select materials, confirm the file name, and view the original content
Parsing configuration	Adjust the processing requirements for this run
Result area	Read and check results after parsing completes

Select Materials Upload a file in the input area, or select a built-in sample. Verify the file name and confirm that the original file can be viewed or played normally.
Set Parsing Requirements
1. In Field Structure, fill in the information you want to extract.
2. Check field names, types, and descriptions as described in Configure Field Extraction.
3. Check the extraction settings and supplementary requirements.
If you already have a suitable configuration, enter testing and reuse it as described in Save and Reuse Configurations.
Submit for Parsing
1. Confirm the selected file and the current settings.
2. Click "Run Document Parsing".
3. Wait for the task to complete. To check processing progress, go to Task Records.
A submitted task uses the settings at submission time. If you change the page settings afterwards, you need to run again to get new results.
View Results and Complete Verification After a text-and-image document completes, you can switch between the Markdown and JSON views:
- Markdown: Best for reading body text, viewing tables, and copying organized content.
- JSON: Best for checking structured content and the information actually returned in the result.
First select a page of the original to compare against, then check headings, paragraph order, table figures, and image descriptions in turn. When you need the result text, use the copy button in the result area. For media materials, check results using segment content and time positions. Video results include four views: "Story Overview", "Parsed Segment Results", "Story Segments", and "Story Summary". We recommend reading the overview or summary first to grasp the overall content, then opening "Parsed Segment Results" and selecting audio or video segments of interest. After clicking a segment with a time position, the player jumps to the corresponding position for easy replay.
Return to Historical Results Go to Task Records, find the corresponding parsing task, view its details, and return to the results from the result entry.