Parseur

Parseur extracts structured data from documents captured through email, an API or direct upload, offering three extraction approaches: Vision AI, Text AI and user built templates. Extracted fields pass through normalisation and validation before being exported to downstream systems.

Capturing from email is the feature that decides whether this fits a given workflow, and it fits a great many. An enormous amount of business data arrives as an email attachment or as the body of a notification, and the usual handling is a person opening each one and retyping it. Giving Parseur an address that receives those messages puts extraction at the point of arrival, with nothing to integrate at the sending end.

Offering three extraction methods rather than one recognises that documents differ. A template is exact and cheap for a layout that never changes, Text AI handles textual documents whose wording varies, and Vision AI reads documents where layout carries meaning. Choosing per document type rather than accepting one approach everywhere is what keeps both accuracy and cost sensible.

Normalisation and validation are the step that makes output usable rather than merely extracted, since a date read correctly but in five different formats still breaks the system receiving it. Pricing starts free and scales by volume.