Does Martini train on user uploads?

Uploaded files — reference images, source video, audio, masks, control inputs — are stored against your project so the workflow can use them. They are sent to the appropriate model provider as part of a generation request, and the result is returned to your canvas. Files are retained while the project exists so you can re-run nodes, branch workflows, and edit. Deleting a project removes the associated uploads on the standard deletion schedule.

Under our Terms of Service, Martini may also use uploads, prompts, generated outputs, and related metadata to operate the platform, run safety and abuse systems, generate analytics and aggregated insights, and train, fine-tune, evaluate, and improve our and our affiliates' AI models, classifiers, and related products. This is how we get better at routing, safety, quality, and feature development over time.

Each AI model on Martini comes from a third-party provider — OpenAI, Anthropic, Google, FAL, Runway, Hailuo, and others. Each provider has its own data and training policy. We route generations through API endpoints rather than consumer chat products, so the provider's API training stance applies, not their consumer-product stance. Where a provider exposes an opt-out from training-on-API-data, our integration uses that opt-out path by default.

If a specific provider has a different default and an opt-out is not available at the API level, that is noted on the model card for the relevant node. Refer to the upstream provider's privacy policy for the exact terms that apply to a specific model — Martini's Privacy Policy governs what Martini does with your data, not what each upstream provider does on its own systems.

We retain the operational data needed to run the platform: your project files and generated outputs (so you can keep working on them), recent prompt and job history (so you can re-run, branch, and audit), credit-spend records (so we can bill correctly), and short-lived job-queue entries during async generation (so video and other long-running jobs can complete and webhook back).

Operational logs are retained for the time needed for debugging, abuse investigation, and billing reconciliation, then purged. Aggregated and de-identified metrics — total generations per model, error rates, queue latency — may be retained longer to monitor platform health and may also be used for product and model improvement, as described in the Privacy Policy.

For high-sensitivity work, check each model card before generating, and delete projects you no longer need from the project menu — uploads and outputs are queued for deletion when you do. Note that data already used to train or evaluate a model, or already incorporated into aggregated or anonymized datasets, may persist in model parameters or derived datasets even after the source content is deleted; we are not required to retrain models in response to a deletion request.

Enterprise and API customers can request additional contractual data-handling terms — including no-training and zero-retention options, bespoke retention windows, regional data routing, and signed processing agreements (DPAs). Reach out via the help icon and ask for an enterprise contact if you need terms beyond the standard policy.