FAQ
Answers to common questions about ParseX capability selection, file limits, configurations, async tasks, result reuse, and billing.
What is the difference between Parse and Extract?
Parse restores files into structured content suitable for Agents and applications, and can process documents/images, audio, and video. Extract pulls strongly typed business fields from document/image content according to a JSON Schema, and can also return field status and source citations.
Can Extract process audio or video?
No. Extract currently only supports document/image files and can only reuse document/image ParseResults. Use Parse for audio and video.
How large a file can I upload on the console playground?
The playground currently limits uploads to 200 MB per file. This limit is not the same as the API limit; the file sizes the API accepts are subject to the official API documentation and service responses. For the complete format list, see Supported Files and Limits.
Why must I run a sample before saving a configuration?
A configuration becomes a stable, reusable asset. Running a real sample and checking the results lets you verify that the parameters, Schema, field status, and citations meet expectations before saving.
Can I pass both config_id and inline parameters?
Yes. You can reference a saved config_id, or pass inline capability settings directly. If both are passed, the inline parameters take precedence.
Why didn't I get results immediately after submitting a task?
Both Parse and Extract are asynchronous tasks. The finalized REST protocol specifies that the submission response returns a biz_id, and the result response returns a status of processing, success, or failed. After the production endpoint is released, final results are only available when the status is success.
How do I troubleshoot a failed task?
After the production REST endpoint is released, callers should use the original biz_id to query the result, and record the request_id, error code, and config_id. Errors in parameters, format, or Schema should be fixed in the request first. For console tasks, check the information provided on the "Task Records" page.
How long can a ParseResult be reused?
Reusable document/image ParseResults are currently retained for 7 days. The result download URL may expire earlier; refer to the expiration time returned by the API.
Does a missing or conflict result from Extract mean the task failed?
No. As long as the system fully evaluated the entire Schema, the REST task status can still be success. missing means the field was not found; conflict means conflicting candidates exist in the source content.
How is ParseX metered?
Usage for Parse and Extract is recorded separately. You can view usage information in the console. For the specific metering rules, prices, quotas, and settlement rules, see the Metering and Billing page.
Why can't I see my configurations or tasks after switching Workspaces?
Files, configurations, tasks, and results are isolated by Workspace. Please switch back to the Workspace you used when creating the resources.