AI Add-on Data Terms
1. What the AI Add-on does
The AI Add-on lets you describe a task in words or attach a file (for example an Excel or CSV list of cartons and pallets) and get a proposal: layers, patterns, a mix, an order, or an explanation of a validation error. The proposal is shown in the Software; nothing is saved until you accept it. Every accepted proposal goes through the same algorithmic checks as work done by hand.
2. What is sent to the model provider
For each request the Software sends:
- the text you typed and the file you attached, as you provided them;
- the products, pallets, layers and settings of the open document that the task needs;
- a fixed description of how the Software works (units, rules, limits), written by us;
- for an error explanation: the error message and the data it refers to.
The Software does not send: station files and station settings beyond the limits needed for the task, your licence key, computer name or other licence data, documents that are not open, or anything from your computer outside the Software.
Do not type or attach personal data (names, e-mails of people) or data you are not allowed to share with a provider outside your company. The provider processes what you send.
3. Who processes it
Standard variant (included allowance). Requests go through our account with Anthropic, the provider of the Claude models. For customers in the European Economic Area, the United Kingdom and Switzerland the provider’s contracting entity is Anthropic Ireland, Limited. Data is processed in the United States under the provider’s Data Processing Addendum with EU standard contractual clauses. The provider does not use your inputs or outputs to train its models (Anthropic Commercial Terms, section B: “Anthropic may not train models on Customer Content from Services”). Inputs and outputs are deleted by the provider within 30 days; content flagged by its safety systems may be kept up to 2 years. We are your processor for this data; we see request counts, token counts and timestamps, not content.
Own-key variant. You enter your own API key from the provider. Requests go straight from the Software to the provider under your own agreement with it; we receive nothing, set no allowance and have no access to the content. Keep the key confidential; the Software stores it only on your computer.
4. Allowance and logging
In the standard variant the allowance is the number of operations per month stated in your order (for example 100). One operation is one request to the model. When the allowance is used up, the function pauses until the first day of the next month. We log for each operation: licence key, time, model, tokens in and out, and whether it succeeded. We do not log the content.
5. Where it does not work
A plant server without internet access cannot use the AI Add-on. The Software shows this instead of sending anything.
6. Proposals are proposals
Language models can be wrong. The Software validates every accepted proposal with the same algorithms as manual work, and the rest is your decision: check the result and test it on the machine before production, as described in the EULA, section 9.
7. Changes
If we change the provider, the models or what is sent, we update these terms and tell you in the Software before the change applies. You can switch the AI Add-on off at any time in the settings.
Contact: contact@palletdesigner.com.