Data privacy during Inference

DeepInfra offers simple, scalable and cost-effective inference APIs. The goal of this document is to explain how we handle data during inference.

Data Privacy

When using DeepInfra inference APIs, you can be sure that your data is safe. We do not store on disk the data you submit to our APIs. We only store it in memory during the inference process. Once the inference is done the is data is deleted from memory.

We also don't store the output of the inference process. Once the inference is done the output is sent back to you and then deleted from memory. Exception to these rules are outputs of Image Generation models which are stored for easy access for a short period of time.

Bulk Inference APIs

When using our bulk inference APIs, you can submit multiple requests in a single API call. This is useful when you have a large number of requests to make. In this case we need to store the data for longer period of time, and we might store it on disk in encrypted form. Once the inference is done and the output is returned to you, the data is deleted from disk and memory after a short period of time.

No Training

The data you submit to our APIs is only used for inference. We do not use it for training our models. We do not store it on disk or use it for any other purpose than the inference process.

No Sharing

We do not share the data you submit to our APIs with any third party.

Logs

We generally don't log the data you submit to our APIs. We only log the metadata that might be useful for debugging purposes, like the request ID, the cost of the inference, the sampling parameters. We reserve the right to look at and log a small portions of requests when necessary for debugging or security purposes.