Mindee
AI & modelsAPI keycredential probe
Document OCR and structured data extraction with Mindee V2 models — invoices, receipts, passports and any other model built from the catalog. Every file route is asynchronous: a file is enqueued against a model and the inference is polled or delivered by webhook.
OCNORA ships 17 typed Mindee operations. Each node has a schema-driven form, input and output tables beside its settings, and is pre-flighted by Workflow Health before your first real run. Anything not covered by a typed node reaches https://api-v2.mindee.net through the HTTP node with the same credential.
Operations 17
- Classify DocumentRun a classification model over a document and return the document type it was assigned.mindee.classify_document
- Crop DocumentRun a crop model over a file and return the cropping coordinates of every document it detects on each page.mindee.crop_document
- Delete RAG DocumentPermanently remove one indexed RAG document from its model.mindee.delete_rag_document
- Extract DocumentRun an extraction model over one invoice, receipt, passport or any other document and return its fields as plain values.mindee.extract_document
- Get Classification ResultFetch a finished classification inference by id and return the document type it assigned.mindee.get_classification_result
- Get Crop ResultFetch a finished crop inference by id and return the regions it detected.mindee.get_crop_result
- Get Extraction ResultFetch a finished extraction inference by id and return its fields as plain values.mindee.get_inference
- Get JobRead the status of one enqueued job, and the URL its result will be readable at once processing finishes.mindee.get_job
- Get OCR ResultFetch a finished OCR inference by id and return its per-page text.mindee.get_ocr_result
- Get RAG DocumentRead one indexed RAG document, including its status and match count.mindee.get_rag_document
- Get Split ResultFetch a finished split inference by id and return the documents it found in the file.mindee.get_split_result
- OCR DocumentRun an OCR model over a document and return the full text of every page, with each word's position.mindee.ocr_document
- Search ModelsSearch the account's models by name and type, which is how a model UUID is found without leaving the workflow.mindee.search_models
- Search RAG DocumentsList the documents indexed against one extraction model for Retrieval-Augmented Generation.mindee.search_rag_documents
- Split DocumentRun a split model over a multi-document file and return the page range and type of each document inside it.mindee.split_document
- Update RAG DocumentActivate or draft one indexed RAG document, and optionally replace its field-level annotation.mindee.update_rag_document
- Upload RAG DocumentIndex one document against an extraction model so later inferences can retrieve it for better accuracy.mindee.upload_rag_document
More in AI & models
Run your first Mindee workflow today.
Create a free workspace, add the connection, pick a template or describe the workflow to Build with AI. Workflow Health checks it before it runs.