wdmpgWhere Do My Prompts Go?
121 sources checked · 0 fetch errors last check data v2026-10-26 · sha256 00e98854 data.jsondata.csvllms.txt

service openai-apivendor OpenAIcategory Model APIs

Does OpenAI API train on your data?

No (Favourable)Not used for training unless you opt in (e.g. via feedback sharing).

fields
5 favourable3 caution
sources cited
3 pages
source monitor
3/3 ok on 2026-10-26
last field check

Evidence log

8 entries · newest check 2026-10-25
OK

Trains on your data by default

= no (Favourable)

Not used for training unless you opt in (e.g. via feedback sharing).

As of March 1, 2023, data sent to the OpenAI API is not used to train or improve OpenAI models (unless you explicitly opt in to share data with us).

WARN

Default retention

= 30 days (Caution)

Abuse monitoring logs up to 30 days. Stateful endpoints (Assistants, Conversations, vector stores, files) keep application state until you delete it.

By default, abuse monitoring logs are generated for all API feature usage and retained for up to 30 days, unless longer retention is required by law, or is reasonably necessary to protect our services or any third party from harm.

Against the 21 tools that state a period (grey dots):

01d7d30d90d1y∞
OK

Zero data retention

= on approval (Favourable)

Zero Data Retention or Modified Abuse Monitoring, per org or project, after OpenAI approves you. Stateful endpoints are not ZDR-eligible.

Currently, these controls are subject to prior approval by OpenAI and acceptance of additional requirements.

WARN

Abuse-monitoring exception

= can override ZDR (Caution)

Even after approval, OpenAI can switch specific models back to retention ('Safety Retention') for severe-risk investigations, with advance written notice.

we reserve the right to make models ineligible for Zero Data Retention or Modified Abuse Monitoring for specific customers if reasonably necessary to investigate or prevent severe risk activity

OK

Data residency / regions

= US, EU, AU, CA, JP, IN, SG, KR, UK, UAE (Favourable)

Per-project regions. Singapore (sg.api.openai.com) and the United Kingdom (gb.api.openai.com) give regional storage only, not regional processing; the same is true for AU, CA, JP, IN and KR. Regional processing is listed only for the US, EU and UAE. Non-US regions require MAM or ZDR approval and a Modified Retention amendment.

To use data residency with any region other than the United States, you must be approved for abuse monitoring controls, and execute a Modified Retention amendment. … United Kingdom* | gb.api.openai.com | Yes | No | Yes

OK

SOC 2 / ISO 27001

= SOC 2 Type 2 (Favourable)

SOC 2 Type 2 for the API Platform. ISO 27001 not confirmed on the pages checked.

Our API Platform has been audited and certified for SOC 2 Type 2 compliance.

OK

HIPAA BAA

= yes (Favourable)

BAA available on request; only BAA-eligible endpoints may process PHI.

We are able to sign Business Associate Agreements (BAA) in support of customers’ compliance with the Health Insurance Portability and Accountability Act (HIPAA).

WARN

Self-host option

= partial (open-weight gpt-oss only) (Caution)

API models are hosted only. OpenAI separately publishes open-weight gpt-oss models (120B and 20B, Apache 2.0) that you can run yourself.

Open reasoning models designed to run locally on desktops, laptops, and in data centers—available in 120B and 20B parameters.

Source checks

3 pages

Spotted something wrong or out of date? Corrections are welcome; see how to send one. How statuses are assigned.

Weekly policy changes. Confirm by email. Unsubscribe anytime.

Powered by Buttondown