How we handle your data.
Four principles we work by.
Minimum data
We ask only for what the task needs and prefer sample or redacted documents at the start.
Agreed locations
Data stays where we agree on it. Your documents are not used to train models unless you agree in writing.
Private options
Open-source models can run inside your environment.
Logging
Runs are logged so results can be audited.
Three places, chosen with you.
| Where | What it means |
|---|---|
| Local or your environment | Open-source models run on infrastructure you control. Data does not leave it. Quality and cost trade-offs are tested first. |
| Cloud model APIs | Data is sent to the model provider agreed for your project, under its API terms. We share the current terms with you before use. |
| Our working storage | Project files are kept with restricted access. API keys and secrets stay out of code and repositories. Logs avoid raw sensitive content where possible. |
NDA first, sample data first.
We sign an NDA first, and a data processing agreement where personal data is involved. We prefer sample or redacted documents at the start.
Deleted after delivery.
We delete your data after final delivery or on your request, and confirm deletion in writing if you ask.
No certifications, no compliance promises.
We hold no security certification such as SOC 2 or ISO 27001, and we give no guarantee of compliance with any regulation. Responsibility for your regulatory obligations stays with you; we build to the constraints you give us.
A written note.
A written data-handling note is available on request.
[ ask for the note → ]