Parse documents

Upload a PDF, Word, Excel, HTML, ODT or RTF file and get back clean, agent-ready content — no URL required. Great for private documents, contracts, reports, and user-uploaded files.

POST/api/v1/parse

Limits & pricing

Maximum file size is 50 MB. Billed at 1 credit base + 1 credit per pageactually processed. A single-page PDF costs 2 credits; a 20-page PDF costs 21.

Request — multipart/form-data (recommended)

Field	Type	Description
filerequired	binary	The document to parse. Up to 50 MB.
options	string (JSON)	Optional JSON string with the same shape as the JSON request body below (formats, pdf, onlyMainContent, timeout).

Request — application/json

Provide either url (we fetch the document server-side) or fileBase64 + filename.

Field	Type	Description
url	string	Absolute URL to a document. We fetch it and infer the filename from the path. Use this OR fileBase64.
fileBase64	string	Base64-encoded file bytes. Use this OR url.
filename	string	Original filename — required with fileBase64; inferred from url otherwise.
contentType	string	Optional MIME type override (e.g. "application/pdf").
options.formats	string[]	Any of "markdown", "html", "rawHtml", "json". Default: `["markdown"]`
options.onlyMainContent	boolean	Strip headers/footers from each page. Default: `true`
options.pdf.mode	"fast" \| "auto" \| "ocr"	Parsing strategy for PDFs. Use "ocr" for scanned documents. Default: `"auto"`
options.pdf.maxPages	number	Cap the number of pages parsed (and credits charged).
options.timeout	number	Hard request timeout in ms. Default: `60000`

Example — upload a PDF

bash

curl https://neuroapi.me/api/v1/parse \
  -H "Authorization: Bearer $NEUROAPI_KEY" \
  -F "file=@./contract.pdf" \
  -F 'options={"formats":["markdown"],"pdf":{"mode":"auto"}}'

Example — parse a remote URL

bash

curl https://neuroapi.me/api/v1/parse \
  -H "Authorization: Bearer $NEUROAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com/whitepaper.pdf",
    "options": { "formats": ["markdown"] }
  }'

Example — JSON / base64

bash

curl https://neuroapi.me/api/v1/parse \
  -H "Authorization: Bearer $NEUROAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "fileBase64": "JVBERi0xLjQK...",
    "filename": "contract.pdf",
    "options": { "formats": ["markdown"], "pdf": { "mode": "auto" } }
  }'

Example response

json

{
  "success": true,
  "data": {
    "markdown": "# Mutual NDA\n\nThis Mutual Non-Disclosure Agreement...",
    "metadata": {
      "filename": "contract.pdf",
      "contentType": "application/pdf",
      "pages": 4
    }
  },
  "creditsCharged": 5
}

Tips

Use pdf.mode: "ocr" for scanned PDFs without a text layer.
Set pdf.maxPages to cap cost on very large documents.
For HTML files, onlyMainContent: true behaves the same way as /scrape.

← Previous

Branding

Interact