ConvertPDF

Current state of a job. Once it succeeded, includes fresh short-lived download URLs for every output. Failed jobs carry an error_code.

job_status

One-step conversion for files the user attached in the chat (ChatGPT passes them as file objects with a download_url). Give the tool slug, the files in the order they should be used, and the tool's options. Downloads the files, runs the tool and returns download links to show the user. Use start_job instead when you can upload files yourself.

convert_files

Delete a job and all of its files now instead of waiting for the automatic one-hour deletion.

delete_job

Call after uploading every file for a job. Seals the uploads, runs the conversion and waits up to about 25 seconds for it. Returns the final status with download URLs, or the current status if it is still running (then call job_status).

finish_job

List every ConvertPDF tool with the file types it accepts, how many files it takes, its options and how many daily units it uses. Call this first to pick a tool slug and options for start_job.

list_tools

Start a conversion: give the tool slug, each file's MIME type and exact size in bytes (in the order they should be used), and the tool's options. Returns a job_id and one signed upload URL per file. Upload each file's raw bytes with HTTP PUT to its url using exactly the given headers (for example: curl -X PUT -H "Content-Type: application/pdf" --data-binary @file.pdf "<url>"), then call finish_job. Never put document contents in tool arguments.

start_job