mode parameter on /parse. Each mode applies a recommended processing configuration, so you don’t have to tune the parser yourself. The parser ships with three:
An agent reviews each segment, flags low-confidence text, and re-reads it against the source to resolve errors at faster speeds. Best for text-heavy documents.
Choosing a Mode
The three modes trade speed for fidelity. The response shape is the same across them in most cases, though jobs submitted withmerge_tables (Agentic Lite and Agentic) can return a cached cross-page merged result instead (see the parse job status reference). Start with Fast and step up only when the output misses detail your pipeline depends on.
Mode Values
Pass the mode’s value in themode field of your /parse request.
You can still send other parameters alongside a mode, such as
merge_tables, validate_segments, extract_charts, or segment_analysis. Anything you send explicitly takes precedence over the mode. See the Parse API reference for the full parameter list.The Agentic mode also digitizes charts with
extract_charts=standard, so Picture segments that are charts come back with chart_data. Send extract_charts=basic to turn that off.Parsing forms? Agentic Lite and Agentic return
KeyValuePair segments, keeping each printed label with its filled-in value. See Element Types.Applying a Mode
The submit-and-poll flow is the same as in the Quickstart; only themode value differs. The snippets below show each mode in Python — translate to JavaScript or cURL by changing the request body shape, not the parameter values.

