Privacy-first isn't a slogan here — it's an architecture. This page is a template for an early-stage product and isn't legal advice.
Last updated · June 2026Account data (name, email, firm) you provide at sign-up; the files you upload and the outputs you generate; and operational logs needed to run and secure the service.
We collect the minimum needed to deliver reconciliation, sign-off, and recurring packs.
Sensitive fields are detected and masked before any data is sent to a language model. The model works against masked, structural information — not raw identifiers.
Calculations run in our own sandboxed engine, so your numbers don't transit a third-party code tool.
To provide the service: profile files, propose cleaning, reconcile totals, generate receipts, and run scheduled refreshes you set up.
We do not sell personal data, and we do not use your client files to train models.
Raw files are stored versioned and immutable so numbers stay traceable to source. Data is scoped to your workspace and isolated from other tenants.
You can delete datasets from your workspace; backups age out on a rolling schedule.
We rely on a small set of infrastructure and model providers to run the service. Each is bound by data-processing terms consistent with our DPA.
Secure-tier deployments can run inside your own environment to narrow this set further.
You can access, export, or delete your data, and request details on processing. Reach us at the security desk on the contact page.
Where applicable law grants additional rights, we honor them.