Data use and privacy

Windpaint does not train on your prompts, inputs or outputs. Who can see your data, where it's stored, and which service providers handle it.

This page covers what Windpaint does with the data you send and the media you generate: your prompts, the files you upload, and the images and clips that come back.

We don’t train on your data

Windpaint does not use your prompts, uploaded inputs or generated outputs to train or fine-tune any model, and never will. This applies to every account, on every plan, whether you use the API, the dashboard, the CLI or the MCP server. There is no setting to opt in or out because there is nothing to opt out of.

We use your data to run the jobs you submit, store the results in your project, bill you for them, and support you when you ask for help.

Who can access your data

Your organization. Everything you upload or generate belongs to a project in your organization. Members of the organization can read it according to their role, and API keys act with their creator’s role or a narrower one. All four roles (owner, admin, member and read_only) can read projects, jobs and assets. See Organizations and projects for the full permission table. People outside your organization can’t list or read your assets through the API.

Signed download links. An asset’s download URL redirects to a signed link that works without a key for 15 minutes. Anyone you hand a signed link to can download that file until it expires. See Assets.

Windpaint staff. Staff can access account, billing and job records, including prompts and outputs, to give support you’ve asked for, to resolve billing issues, and to investigate abuse or a breach of our terms. We don’t browse customer content for any other reason.

How long we keep it

Uploaded files and generated outputs are kept for your organization’s retention period, 7 days by default and configurable from 1 to 365 days or forever. When an asset expires or you delete it, its file is deleted from storage within minutes. The record of the asset and the job that made it, including the prompt, stays in your project. See Retention and storage.

When you pass an https:// URL as a job input, Windpaint downloads the file from that URL and stores it as an asset in your project, under the same retention.

Content moderation

Prompts, and in some cases input images and generated images, are checked against Windpaint’s content policy before a job runs and after it finishes. When something is blocked, Windpaint records the event against your organization, with the categories that matched, so repeated serious violations can be reviewed by staff.

Where your data is stored

Where models run

Both live models, z-image-turbo and flux-2-klein-4b, are open-weight models (Apache-2.0) that Windpaint runs on its own GPUs. To run a job, your prompt and any input files are sent to those GPUs, and the output comes back to storage in us-east-1.

Your prompts and inputs are not sent to a third-party model API to generate the live models’ outputs. If we add a model that runs on another provider’s API, we’ll say so on the Models page and update the list below. Whatever provider runs a model, Windpaint’s own commitment above stays the same: we don’t train on your data.

Service providers

These are the third parties that process customer data on Windpaint’s behalf. Current as of 2026-10-05.

Questions

If you have a question about how your data is handled, contact support.