Documents come into Pactus, are stored encrypted and are read by Pactus's own OCR. To answer questions it uses a language model. You choose where each part runs: in our EU cloud, in your own infrastructure, or entirely on your own servers.
It's the same app every time. What changes is where it runs and where the models run.
We run Pactus, OCR included, for you in EU data centres. Your documents are stored encrypted, and the language model it calls is EU-based and keeps nothing once it has answered.
Pactus, OCR included, runs in your cloud or on your servers, so your documents are stored with you and you stay the only data controller. For the language model it calls an EU-based cloud model with zero data retention.
Pactus and the language model run on your own hardware, with or without internet access. Nothing leaves your network and we have no access to your documents.
Documents are encrypted while they travel to Pactus and while they are stored.
Neither we nor the model providers train any model on your documents.
Each customer's data is processed and stored in its own environment.
Deleting a document removes it from storage, along with everything Pactus extracted from it.
You set who sees what, per workspace and per case. Pactus answers each person only from the documents they are allowed to open.
Opening a document, asking a question and changing a result are all recorded, and your administrators can review the log.
When we host Pactus, we sign a data processing agreement with you. On your own infrastructure, you remain the data controller.
We'll walk your IT and compliance team through the setup and send the documentation your procurement needs.