How Due Badger uses AI

How Due Badger uses AI.

Due Badger is built to be driven by an AI assistant, which makes it worth being exact about where a model is involved and where one is not.

There is no model inside Due Badger today

Nothing in the running service sends your register to a language model. Reminders are worked out by date arithmetic. The escalation ladder is a fixed set of rules. No part of the product writes text about your obligations.

Your assistant is yours

Due Badger is an MCP server, so an assistant you already use, such as Claude, can read and update your register on your behalf. That assistant is your relationship and not mine. What it does with what it reads is governed by whoever provides it, and your register is only ever reached with an OAuth token issued to your account.

What an agent can and cannot do

The tools are the same rules the web app follows. There is no tool that deletes anything. Closing an occurrence requires a note, and moving a date requires a written reason, at the signature of the tool, so no client can skip either. Archiving is reversible.

Training

I do not use your register to train any model, and I do not sell or share it. See the Privacy notice for what happens to your data generally.

The monthly regulatory scan

The scan described on the home page is the one planned feature that will use a model: it reads public regulator sources for the frameworks you carry and reports what changed. It is not built yet, and it is on the roadmap rather than in the product. When it does ship, every finding will carry the source it came from, nothing will be added to your register without you accepting it, and your own register will not be sent anywhere to produce it.