file content part and Nito parses it into text before the model runs. Supported: PDF, DOCX, PPTX, XLSX, XLS, and plain text. Images take a different path: Image Input.
The File Part
file_url: anhttpsURL with no credentials and no private or loopback host, orfile_data: the bytes inline, withfilename: helps the parser identify the kind.
file and input_file are both accepted as the part type. There is no file_id; supply the document on each request.
Limits
A document that hits a limit fails with a sanitized error rather than returning a shortened document. Split very large PDFs first.
Related
Multimodal
Overview.
File Parser
How parsing works.
Image Input
Images in chat.
How the API Works
The request lifecycle.
Extract text without inference
Upload a multipartfile to POST /api/v1/attachments/parse. The API returns extracted text, a model-independent token estimate, and the document type. No model ID is required; document parsing is billed separately.
