Understanding Data Storage and Retrieval in OpenAI's ChatGPT API

API Bugs
, ,
1

Hello everyone,

I’ve been working with OpenAI’s ChatGPT API for a while now and I have a couple of questions regarding data storage and retrieval:

  1. Data Storage:
  • When we use the ChatGPT API, is our data stored by OpenAI? If so, for how long?
  • What are OpenAI’s policies regarding the use of our data? Is it used to improve their models or for any other purpose?
  1. Retrieving Previous Requests:
  • Is there a way to retrieve previous API requests and responses? Does the API support session tracking or is it completely stateless?
  • If the API doesn’t provide a built-in way to access previous interactions, what are some best practices for logging and storing conversation data on the client side?

Any insights or examples on how to effectively manage and log conversation history while ensuring data privacy and security would be greatly appreciated.

Thanks in advance for your help!

2

Hi @anjula!

To answer a few of your queries:

Data processed via the API is by default NOT used for training purposes by OpenAI unless you specifically opt in for this. However, data is generally stored for a period of 30 days. There’s the possibility to apply for zero data retention - however, we do not have any good data points so far from the Forum how often this has been granted in practice. You can read up details on OpenAI’s policy for data handling here.

For the core API there is no log of your requests that you can access. Exception are API calls performed via the Playground where you can see the history of your requests.

1 Like