https://ainception.ca/blog/ Turning document text into structured JSON: notes on the DocExtract API and document-extraction practice. en-ca Fri, 18 Sep 2026 19:19:18 +0000 https://ainception.ca/blog/docextract-business-mega-correction-normalization/ https://ainception.ca/blog/docextract-business-mega-correction-normalization/ Fri, 18 Sep 2026 00:00:00 +0000 OCR-style correction, field-specific date and numeric formats, literal text substitution, and a change-by-change audit. What the Mega plan adds, with verified postprocessing examples and its limits. https://ainception.ca/blog/contract-text-to-json-api/ https://ainception.ca/blog/contract-text-to-json-api/ Thu, 10 Sep 2026 00:00:00 +0000 You already have the contract as text. Turn it into typed JSON — parties and roles, effective and expiration dates, contract value, governing law — with one API call. No OCR. https://ainception.ca/blog/extract-json-from-pdf-text-layer-without-ocr/ https://ainception.ca/blog/extract-json-from-pdf-text-layer-without-ocr/ Thu, 10 Sep 2026 00:00:00 +0000 Most generated PDFs already carry a real text layer. Pull it with pdftotext or pdfminer.six, then turn that text into typed JSON with one API call — no OCR, no per-page pricing. https://ainception.ca/blog/extraction-api-pricing-flat-vs-per-page/ https://ainception.ca/blog/extraction-api-pricing-flat-vs-per-page/ Thu, 10 Sep 2026 00:00:00 +0000 Per-page extraction pricing punishes long documents. If you already have the text, a flat monthly call quota is simpler to forecast. Here is the trade-off, with the real numbers and a verified API response. https://ainception.ca/blog/general-document-type-messy-documents/ https://ainception.ca/blog/general-document-type-messy-documents/ Thu, 10 Sep 2026 00:00:00 +0000 Most real-world text fits no preset schema. DocExtract's general document_type gives a best-effort structured summary of any text. Here is what it returns, with two verified API responses, and where it stops. https://ainception.ca/blog/meeting-notes-text-to-action-items-json/ https://ainception.ca/blog/meeting-notes-text-to-action-items-json/ Thu, 10 Sep 2026 00:00:00 +0000 You already have the meeting notes or transcript as text. Turn it into typed JSON — action items with owners and due dates, decisions, attendees — with one API call. No OCR. https://ainception.ca/blog/ocr-output-to-structured-json-api/ https://ainception.ca/blog/ocr-output-to-structured-json-api/ Thu, 10 Sep 2026 00:00:00 +0000 You have already run Textract, Document AI, or Tesseract. Take the recognised text — noise and all — and turn it into typed JSON with one API call. No templates, flat monthly quota. https://ainception.ca/blog/resume-text-to-json-api-no-ocr/ https://ainception.ca/blog/resume-text-to-json-api-no-ocr/ Thu, 10 Sep 2026 00:00:00 +0000 If you already have the resume text — pasted, from an email, or a text-layer PDF — you can skip the upload widget and OCR. Turn CV text into typed JSON with one API call. https://ainception.ca/blog/document-extraction-api-vs-diy-vs-l1m/ https://ainception.ca/blog/document-extraction-api-vs-diy-vs-l1m/ Wed, 09 Sep 2026 00:00:00 +0000 You can already get JSON out of GPT or Gemini. Here is an honest look at when a narrow extraction API is worth it instead — and when it is not. https://ainception.ca/blog/parsing-invoices-to-json-without-ocr/ https://ainception.ca/blog/parsing-invoices-to-json-without-ocr/ Wed, 09 Sep 2026 00:00:00 +0000 If your invoices already arrive as text or text-layer PDFs, you can skip OCR entirely. How to turn invoice text into typed JSON with one API call. https://ainception.ca/blog/document-text-to-structured-json-api/ https://ainception.ca/blog/document-text-to-structured-json-api/ Tue, 08 Sep 2026 00:00:00 +0000 How to convert invoice, resume, contract, and meeting-note text into typed JSON with a single API call. Free tier, no OCR setup, no document storage.