API reference

Three extraction endpoints, one response shape. Base URL for all requests is https://www.procuremind.io. New here? Start with the developer overview.

Authentication

POST /api/extract accepts an API key in the X-API-Key header (Pro/Team plans), a Supabase session token as Authorization: Bearer, or a dashboard cookie session.

The file endpoints (/api/extract/pdf, /api/extract/excel, /api/extract/docx, /api/extract/image) currently authenticate with Authorization: Bearer or a cookie session only. API-key auth is not yet enabled on file uploads. File upload also requires a paid plan.

All responses are JSON. Errors return { "error": "..." } with the status codes listed per endpoint.

POST`/api/extract`Extract structured data from a quote email's text: the developer-API endpoint.

Request

Content-Type: application/json · Auth: X-API-Key (Pro/Team), Bearer token, or cookie session

Field	Type	Description
`email_body`	string	Required. The plain-text body of the quote email. Maximum 100 KB.
`email_subject`	string	Optional. Email subject line: improves document-type and quote-number detection.
`email_sender`	string	Optional. Sender address: used for vendor detection and duplicate matching.

Response: 200 OK

Field	Type	Description
`extraction_id`	string (uuid)	ID of the saved extraction. Reference it in support requests.
`extracted_data`	object	The extracted fields (see the extracted_data schema below).
`confidence_score`	number	0–1 confidence score (same value as extracted_data.confidence_score).
`model_used`	string	AI model that produced the extraction (e.g. gpt-4o-mini).
`tokens_used`	number	Tokens consumed by the extraction. Absent on cached responses.
`cached`	boolean	Present and true when the same subject + sender + body was extracted before. The stored result is returned without consuming quota.

Example

curl -X POST https://www.procuremind.io/api/extract \
  -H "Content-Type: application/json" \
  -H "X-API-Key: pm_your_key_here" \
  -d '{
    "email_subject": "Quote Q-1042 - Acme Fasteners",
    "email_sender": "sales@acme-fasteners.com",
    "email_body": "Hi, please find our quote: 500x M6 hex bolts @ $0.12 each..."
  }'

Errors

Status	When
`400`	Invalid JSON body, email_body missing, or body larger than 100 KB.
`401`	No valid API key, Bearer token, or session.
`402`	Monthly extraction limit reached. Response includes limit, current_count and upgrade_url.
`403`	Valid API key on a Free or Starter plan. API access requires Pro or Team.
`429`	Rate limit exceeded (10/min per user). Retry after the number of seconds in the Retry-After header.
`500`	Extraction or persistence failure; safe to retry.

POST`/api/extract/pdf`Extract structured data from a PDF quote attachment.

Request

Content-Type: multipart/form-data · Auth: Authorization: Bearer, X-API-Key (Pro/Team) or cookie session · Paid plans only

Field	Type	Description
`file`	file	Required. The PDF document, sent as the multipart field named "file". Must be application/pdf, under 10 MB.

Text-based PDFs are parsed directly; scanned PDFs fall back to OCR, and low-confidence results are retried with vision extraction in batches of 5 pages, merged into a single result. Long text documents are chunked automatically with the same merge. The source field in the response tells you which path produced the result.

Scanned-document page caps per plan (vision OCR):

Plan	Max pages per scanned document
Free	5 (file upload itself requires a paid plan)
Starter	10
Pro	50
Team	100

Response: 200 OK

Same fields as POST /api/extract, plus:

Field	Type	Description
`source`	string	Extraction path: pdf_text, pdf_ocr, or pdf_vision.
`page_count`	number	Number of pages read from the PDF.
`pages_processed`	number	Present on pdf_vision results: pages actually processed by vision (equals page_count unless processing stopped early).
`partial`	boolean	Present and true when the time budget stopped chunked processing early — the result covers pages_processed of page_count pages.

Example

curl -X POST https://www.procuremind.io/api/extract/pdf \
  -H "Authorization: Bearer <session token>" \
  -F "file=@quote.pdf;type=application/pdf"

Errors

Status	When
`400`	No file field, file is not application/pdf, or the file fails PDF validation.
`401`	No valid Bearer token or session.
`402`	Free plan (PDF upload requires a paid plan), or monthly extraction limit reached.
`413`	File larger than 10 MB.
`422`	PDF could not be read or contains no readable text — or the document exceeds your plan's page cap for scanned documents (response includes pages, page_cap and upgrade_url).
`429`	Rate limit exceeded (3/min per user). See the Retry-After header.
`500`	Extraction failure; safe to retry.
`504`	OCR exceeded the processing time limit; try a text-based PDF or fewer pages.

POST`/api/extract/excel`Extract structured data from an Excel or CSV quote file.

Request

Content-Type: multipart/form-data · Auth: Authorization: Bearer, X-API-Key (Pro/Team) or cookie session · Paid plans only

Field	Type	Description
`file`	file	Required. The spreadsheet, sent as the multipart field named "file". Supported: .xlsx, .xls, .csv, under 10 MB.

Response: 200 OK

Same fields as POST /api/extract, plus:

Field	Type	Description
`source`	string	Always excel.
`sheet_count`	number	Number of sheets read from the workbook.
`row_count`	number	Number of rows read.

Example

curl -X POST https://www.procuremind.io/api/extract/excel \
  -H "Authorization: Bearer <session token>" \
  -F "file=@quotes.xlsx"

Errors

Status	When
`400`	No file field, unsupported file type (only .xlsx, .xls, .csv), or the file fails validation.
`401`	No valid Bearer token or session.
`402`	Free plan (file upload requires a paid plan), or monthly extraction limit reached.
`413`	File larger than 10 MB.
`422`	The file could not be read as a spreadsheet.
`429`	Rate limit exceeded (3/min per user). See the Retry-After header.
`500`	Extraction failure; safe to retry.

POST`/api/extract/docx`Extract structured data from a Word (.docx) quote document.

Request

Content-Type: multipart/form-data · Auth: Authorization: Bearer, X-API-Key (Pro/Team) or cookie session · Paid plans only

Field	Type	Description
`file`	file	Required. The Word document, sent as the multipart field named "file". Only .docx, under 10 MB — validated by extension and by content (a .docx is a ZIP container, checked by signature).

Text is extracted from the document and run through the same pipeline as email extraction — document classification, fields and confidence scoring are identical.

Response: 200 OK

Same fields as POST /api/extract, plus:

Field	Type	Description
`source`	string	Always docx.

Example

curl -X POST https://www.procuremind.io/api/extract/docx \
  -H "Authorization: Bearer <session token>" \
  -F "file=@quote.docx"

Errors

Status	When
`400`	No file field, file is not a .docx, or the file fails validation.
`401`	No valid Bearer token or session.
`402`	Free plan (file upload requires a paid plan), or monthly extraction limit reached.
`413`	File larger than 10 MB.
`422`	The document could not be read, or contains no readable text.
`429`	Rate limit exceeded (3/min per user). See the Retry-After header.
`500`	Extraction failure; safe to retry.

POST`/api/extract/image`Extract structured data from a photo or screenshot of a quote.

Request

Content-Type: multipart/form-data · Auth: Authorization: Bearer, X-API-Key (Pro/Team) or cookie session · Paid plans only

Field	Type	Description
`file`	file	Required. The image, sent as the multipart field named "file". Supported: .png, .jpg, .jpeg, .webp, under 10 MB. The real format is detected from the file content (magic bytes), not the extension.

Images go straight to vision extraction — one image is one extraction, counted against the same monthly quota as everything else.

Response: 200 OK

Same fields as POST /api/extract, plus:

Field	Type	Description
`source`	string	Always image.

Example

curl -X POST https://www.procuremind.io/api/extract/image \
  -H "Authorization: Bearer <session token>" \
  -F "file=@quote.jpg"

Errors

Status	When
`400`	No file field, unsupported extension (only .png, .jpg, .jpeg, .webp), or the file is not a valid image.
`401`	No valid Bearer token or session.
`402`	Free plan (file upload requires a paid plan), or monthly extraction limit reached.
`413`	File larger than 10 MB.
`429`	Rate limit exceeded (3/min per user). See the Retry-After header.
`500`	Extraction failure; safe to retry.

The `extracted_data` object

All three endpoints return the same schema. Fields that don't apply to a document type come back as empty strings.

Field	Type	Description
`document_type`	string	One of quote, po, bom, rfi. Defaults to quote.
`vendor_name`	string	Supplier name.
`vendor_email`	string	Supplier email address.
`buyer_name`	string	Issuing organisation (PO, BOM, RFI; blank for quotes).
`buyer_email`	string	Buyer email (PO, BOM, RFI; blank for quotes).
`delivery_address`	string	Ship-to address (PO only).
`quote_number`	string	Document reference: quote number, PO number, BOM ID or RFI reference, depending on document_type.
`bom_version`	string	BOM revision, e.g. "Rev B" (BOM only).
`currency`	string	Document currency code.
`total_amount`	string	Grand total, as a string.
`subtotal`	string	Pre-tax subtotal.
`tax_amount`	string	Tax amount (e.g. GST/VAT).
`shipping_amount`	string	Shipping/freight amount.
`due_date`	string	Due or delivery date.
`quote_validity`	string	Quote/RFI response deadline (blank for BOM/PO).
`payment_terms`	string	Payment terms (quote and PO only).
`incoterms`	string	Incoterms (quote and PO only).
`items`	LineItem[]	Line items (see the LineItem table below).
`confidence_score`	number	0–1 confidence score for the extraction.

`items[]`: the LineItem object

Field	Type	Description
`description`	string	Item description.
`sku`	string	Part number / SKU / product code.
`quantity`	string	Quantity ordered/quoted.
`unit`	string	Unit of measure (each, kg, m, box…).
`unit_price`	string	Price per unit.
`line_total`	string	Extended line total.
`currency`	string	Per-line currency (may differ from document currency).
`discount`	string	Discount, e.g. "10%" or "50.00".
`moq`	string	Minimum order quantity.
`spq`	string	Standard pack quantity.
`lead_time`	string	Lead time, e.g. "4 weeks".
`manufacturer`	string	Manufacturer name (BOM/PO extended fields).
`manufacturer_pn`	string	Manufacturer part number (MPN).
`bom_level`	string	Hierarchy depth in a BOM, e.g. "1.1".
`reference_designator`	string	PCB/schematic references, e.g. "R1, R2, C3" (BOM).
`notes`	string	Line-level notes or RFI question-response text.

Webhooks

Pro and Team accounts can register webhook endpoints in the dashboard under Settings → Webhooks. After every successful extraction, ProcureMind POSTs a signed extraction.completed event to each active endpoint.

Each endpoint gets a signing secret (whsec_…) shown once at creation. Verify deliveries by recomputing HMAC-SHA256(secret, "{timestamp}.{raw body}") and comparing it to the v1 value in the X-ProcureMind.io-Signature header. Delivery times out after 10 seconds and is fire-and-forget. Webhook failures never affect the extraction response.

POST https://your-server.example/hooks/procuremind
Content-Type: application/json
X-ProcureMind.io-Event: extraction.completed
X-ProcureMind.io-Signature: t=1756723200,v1=9f86d081...

{
  "event": "extraction.completed",
  "created_at": "2026-09-01T12:00:00.000Z",
  "data": {
    "extraction_id": "9f1c2a7e-...",
    "document_type": "quote",
    "vendor_name": "Acme Fasteners",
    "total_amount": "60.00",
    "currency": "USD",
    "confidence_score": 0.97,
    "source": "email",
    "sent_to_sheet": false
  }
}

Questions?

Email support@procuremind.io and include the extraction_id: we can trace the request exactly.

Create an account →

API reference

Authentication

POST/api/extractExtract structured data from a quote email's text: the developer-API endpoint.

Request

Response: 200 OK

Example

Errors

POST/api/extract/pdfExtract structured data from a PDF quote attachment.

Request

Response: 200 OK

Example

Errors

POST/api/extract/excelExtract structured data from an Excel or CSV quote file.

Request

Response: 200 OK

Example

Errors

POST/api/extract/docxExtract structured data from a Word (.docx) quote document.

Request

Response: 200 OK

Example

Errors

POST/api/extract/imageExtract structured data from a photo or screenshot of a quote.

Request

Response: 200 OK

Example

Errors

The extracted_data object

items[]: the LineItem object

Webhooks

Questions?

POST`/api/extract`Extract structured data from a quote email's text: the developer-API endpoint.

POST`/api/extract/pdf`Extract structured data from a PDF quote attachment.

POST`/api/extract/excel`Extract structured data from an Excel or CSV quote file.

POST`/api/extract/docx`Extract structured data from a Word (.docx) quote document.

POST`/api/extract/image`Extract structured data from a photo or screenshot of a quote.

The `extracted_data` object

`items[]`: the LineItem object