ade parse to turn a document into Markdown and a structured element tree, saved on your machine. Parse is the first step in most workflows: ade find, ade crop, and ade view all read a completed parse. Run Extract on a completed parse as well: that is what grounds each extracted value to a location on the page. Extract also accepts Markdown from other sources. The CLI covers Parse and Extract, so Classify, Split, and Section stay API-only.
For a worked example, see the Quickstart. For the files a run saves, see CLI Concepts.
Choose Your Input
Parse a local file with-d, or a URL the server fetches with --document-url. Pass exactly one of them. Parse accepts PDFs and images, and rejects encrypted PDFs, so decrypt those first. See File Types.
A URL parse never gives the CLI the document bytes, so ade view and ade crop fetch a copy the first time you use them. Add --keep-copy to fetch it during the parse instead, which matters when a pre-signed URL may expire before then.
Choose a Service Tier
Parse and Extract use async jobs, and each submitted job sends a service tier. The CLI defaults topriority, the fast lane, which consumes credits at the full rate. The standard tier runs in a slower lane and consumes half the credits.
That default differs from the API, where a job that sends no service_tier runs on standard. If you are moving a workload from the API to the CLI, pass --tier standard to keep consumption the same.
For the rates, see Credit Consumption. For how the two lanes differ, see Sync vs Async.
Set the tier per run:
Parse Only Some Pages
Pass 1-indexed page numbers and ranges to--pages. Parsing fewer pages consumes fewer credits.
Send Advanced Options
The--options flag passes a JSON object straight through to the API, for settings the dedicated flags do not cover, such as table format or inline Markdown. The server rejects unknown keys. For the full set, see Parse Input and the CLI Reference.
Parse a Long Document
A document with more pages than your per-minutepriority limit cannot run on the priority tier at all, because all priority work shares that limit. Run it as a standard job instead. See Rate Limits.
A long parse can also outlast the 600-second default wait. The job keeps running on the server, and re-running the same command reconnects to it rather than starting over. CLI Concepts covers the wait budget, and Scripting and Automation shows how to submit now and collect later.
Next Steps
Extract Fields
Run a schema against a parse and read the grounded result.
Find, Crop, and View
Search a parse locally and turn matches into visual evidence.