Security

Enterprise-Grade Security, By Design

Your data is yours — Kruncher processes it on your behalf only. Each customer has a dedicated, isolated tenant. All data is encrypted and never exposed to the public internet. We are ISO 27001 certified, SOC 2 Type 2 certified, and fully GDPR compliant. Your data will never be used for model training, following standards used by leading banks. Kruncher supports many different models and is a partner of Anthropic, NVIDIA, Microsoft Azure and Google, and the platform can be deployed in your own private cloud so data residency stays under your control.

Security

Enterprise-grade security you can trust

Built for private-market workflows: strict access controls, encrypted storage, and audited processes.

Private by Design

Segregated data
Zero-trust access
Zero data retention (ZDR)
Never used for AI training

Protected at Every Layer

End-to-end encryption
Secure infrastructure
Continuous monitoring
Deployment on Private Cloud available

Certified & Compliant

ISO 27001 certified
SOC 2 Type 2 Certified
GDPR compliant
Third-party audits

Control & Data Visibility

Control data access
100% visibility
Model orchestration and sovereign AI available on demand

Access and control

A place your team is allowed to do this work

A CIM, a banker deck or a set of meeting notes cannot go into a consumer chatbot, so the question is not whether AI is useful but where it is permitted to run. Kruncher is built to be that place: your material stays in your tenant, roles decide who sees what, and every answer carries the source behind it.

Who sees what

Roles decide what each teammate can see, set per person on your Team Members screen
Single sign-on available on the enterprise plan
Each customer in a dedicated, isolated tenant, encrypted and never exposed to the public internet

What we do with your material

Your documents and derived data are never used to train AI models
Your diligence method, scoring and corrections stay inside your tenant, with no shared learning across customers
Every figure carries the source and date behind it, so an answer can be checked rather than trusted

Models and deployment

Your choice of model, running where your data has to stay

Kruncher is model agnostic by design and supports many different models. We are partners of Anthropic, NVIDIA, Microsoft Azure and Google, and the whole platform can be deployed inside your own private cloud so that data residency stays under your control.

Anthropic

Claude models across analysis and extraction, and the Model Context Protocol Kruncher exposes its data layer through.

NVIDIA

Accelerated compute behind document ingestion, extraction and the analysis pipeline.

Microsoft Azure

Isolated tenants governed by Microsoft Entra, and the deployment target most institutions already run on.

Google

Gemini models, and Google Cloud as a deployment target for teams standardized there.

Model flexibility

Multiple model providers, selected per task rather than fixed to one vendor
Runs against the models your risk function has already approved
Open source and self hosted models supported
Model orchestration and sovereign AI available on demand

Private cloud and data residency

Deployable into your own AWS, Azure or GCP, inside your network boundary
Documents and derived data never leave that boundary
No egress to a vendor endpoint required, with compute on your account
Available as SaaS where standard procurement is preferred

Legal & policies

Our terms and policies

The agreements and policies that govern how you use Kruncher and how we handle your data.

Terms and Conditions

The terms that govern your use of the Kruncher platform and website.

How we collect, use, store, and protect personal data — and the rights you have over it.

MSA

Our Master Services Agreement — the contract covering enterprise use, data handling, and service commitments.

Talk to the team

Curious how Kruncher could work for your firm?

Meet our data & AI experts: we’ll walk through your workflow and data sources.