What you’ll have at the end: one file processed by fileScout and fileForensics, with its extracted fields, forensic checks and text in front of you, and a sense of what the API will return.
You need: a Developer Portal account. Sign up or sign in. New accounts start with $7 of free credit; see Billing.
The Developer Portal works with Developer Portal workspaces. If your workspace was set up in the fileAI app, create keys in the app instead: Settings → API Keys at orion.file.ai. To try fileAI in the app, follow the App quickstart.
1

Choose your products

On the portal Home page, under Try your first document, choose fileScout to extract data, fileForensics to check for tampering, or both.
2

Upload a file

Drag a file onto the upload area, or click it to browse. You can upload JPG, JPEG, PNG, TIF, TIFF, HEIC, HEIF, PDF, CSV, XLSX, XLS, DOC, DOCX, ZIP and WEBP files, among others; the full list is on Supported file types. To add more files later, open Sandbox in the sidebar and click Upload files.
3

Wait for processing

Sandbox lists every file with the products it used, its schema and its processing status. Open the file once its status is Completed.
4

Read the results

The file opens beside three tabs. Each shows a readable Preview and the raw JSON:
  • Schema: the fields fileScout extracted. Click Verify schema when the values are right, so later files of the same type are filled the same way.
  • Forensics: an overall result, then every check by group, each with its status: Nothing found (none), Issue (issue), or found, skip or error. Open a group to see what each check found. See Detector status.
  • OCR: the text read from each page.
5

Check the cost

Open Usage to see the run, its pages and what it cost. A 6-page PDF with both products costs 6 × $0.015 + $0.03 = $0.12.

​
Next steps

API quickstart

Do the same from code with an API key.

Connect an agent

Ask Claude, Codex or Cursor about your files.

fileScout

How fileScout reads each element of a file.

fileForensics

All 55 checks and what they report.