The @wetransform/core npm package is the foundation of WeTransform, an AI-driven file import layer designed to sit inside B2B SaaS products. Developers integrate it into their React, Vue, or vanilla JavaScript application in approximately four lines, exposing a modal or inline interface to end customers. An iframe alternative, complete with custom-domain hosting, exists for teams that prefer not to ship additional client-side code.
Where traditional file uploaders stop after the upload event, WeTransform continues the pipeline. Uploaded files are parsed regardless of format (CSV, Excel, PDF, XML, JSON), mapped against a target schema configured once by the SaaS team, validated against business rules and constraints, and delivered to the application backend via REST or webhook.
Several AI capabilities power this flow. AI Mapping matches incoming columns to schema fields with confidence scoring. Autoclean handles unstructured material like PDFs containing tables or Excel files with merged cells. Fill the Gaps reconstructs missing values by reading row-level context. Natural language rules let business users describe transformations in plain English or French. Power Functions invoke external services (REST endpoints) mid-transformation, enabling enrichment such as SIRET verification, address geocoding, or product identifier lookups.
Ingestion is multi-channel. Beyond browser uploads, files can arrive via REST API, webhook trigger, FTP, SFTP, URL polling, or dedicated email mailboxes.
Adopters include Cargoo, Sellermania, Autocab, ING, 23 Prod, Veloris, Senef, and Armée du Salut, spanning logistics, marketplaces, transportation, banking, e-commerce, and non-profit operations. Onboarding accelerates from weeks to minutes, engineering recovers roughly 15 percent o