Security
Security and data handling
Your models, drawings and standards are the core of your business. Ideatura treats them as confidential, processes them only to run the work you ask for, and never uses them to train any model. Here is exactly how that works.
Our commitments
What we commit to
Your data is not used for training
We do not use your project data, drawings, models or workflows to train, refine or improve any AI model. Every AI provider we route through is bound by no-training terms, listed below.
Your files and models stay yours
Your project data, drawings, templates, naming conventions and production methodologies remain your property. We process them only to provide the service, and never to train models.
AI providers
The AI providers we use
Different parts of the platform run on different models. Here is every provider we send requests to, the models we use from each, and the terms that keep your data out of their training.
OpenAI's API data policy states that data sent to the API is not used to train or improve OpenAI's models.
Anthropic's commercial terms state that content sent through the API is not used to train Anthropic's models.
The GCP Service Terms state that Google does not use customer data to train or fine-tune its models.
Cohere's enterprise terms let customers opt out of model training, and we have opted out.
OpenRouter's terms leave training control up to each provider, so we disable providers that train on customer content and route only to those confirmed not to.
Questions?
Need this for your DPO or IT team?
We share full security and data-processing documentation with every pilot customer.