Automatically identify what kind of document you uploaded — invoice, contract, resume, report — with key fields detected and suggested next-step tools.
Pro AI tools send document content to xAI.
AI Classifier is a Pro feature — included in PDFwarp Pro at $9 USD/month along with every other tool and unlimited conversions.
The first 5,000 characters of extracted text go to xAI with a request for JSON: type, confidence (High/Medium/Low), a one-sentence summary, key_fields[], and suggested_actions[] (PDFwarp tools). The page renders that object. It is a triage hint, not a records-management system, and it does not store a filing cabinet of your uploads. Pro only.
Use the type and suggested tools to decide whether the next click is AI Extract, AI Contract, or just file it. Do not auto-route money or legal work from Medium/Low confidence.
Enough to decide if you have the right document. Not enough to skip reading a contract.
A letter that is also a memo should come back uncertain. That is a useful 'open this yourself' signal, not a bug.
user_msg uses text[:5000]. The model is asked for JSON with those five keys and no markdown. A parse failure becomes {type: Unknown, confidence: Low, summary: <raw>}.
The prompt asks for PDFwarp tool suggestions. They are labels in the JSON, not deep links we execute. You still pick the next tool.
Classification is not written to a document database. Same xAI + temp-cleanup rules as the rest of the AI suite.
Whatever label the model picks — invoice, contract, resume, letter, report, form, receipt, and others. There is no closed enum we guarantee.
type, confidence, summary, key_fields, suggested_actions. Formality is not a separate field in the API response.
One upload per run. There is no bulk endpoint. Merge only if you want one label for the combined packet — usually you do not.
AI Classifier requires a Pro account. Upgrade at pdfwarp.com/pricing.