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

service together-aivendor Together Computercategory Inference hosts & routers

Does Together AI train on your data?

No (Favourable)Only with explicit opt-in.

fields
5 favourable2 unfavourable1 unverified
sources cited
6 pages
source monitor
6/6 ok on 2026-10-26
last field check

Evidence log

8 entries · newest check 2026-10-21
OK

Trains on your data by default

= no (Favourable)

Only with explicit opt-in.

We do not use any data collected from you to train our models without your explicit opt-in and consent.

srctogether.ai/privacy ↗
FAIL

Default retention

= stored until deleted (default) (Unfavourable)

By default Together stores prompts and responses and may use them for product improvement (not training). You can delete them, or switch storage off to get ZDR going forward.

Unless your organization enables ZDR, Together stores the prompts you send and the responses models return, and may use them for product improvements.

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

01d7d30d90d1y∞
OK

Zero data retention

= yes (opt-in, org setting) (Favourable)

Off by default. Admins set 'Store prompts and model responses' to No in Organization Settings; this also disables passthrough models. Not retroactive.

ZDR is not enabled by default.

OK

Abuse-monitoring exception

= none stated for content (uses metadata) (Favourable)

The terms say abuse monitoring uses Usage Data (logs, metrics, metadata), which excludes your content. No separate abuse-monitoring retention of prompts is stated. Without ZDR, prompts are stored by default anyway (see Retention).

Usage Data may include information such as system logs, performance metrics, model-usage statistics, metadata, and other operational information but excludes Your Content or any personal information. We may use the Usage Data for purposes including improving, maintaining, and securing the Services, developing new features or services, monitoring for abuse

UNVERIF

Data residency / regions

= unverified (Unverified)

No region selection documented for serverless inference. The GPU clusters page says Asia and Middle East options are 'available based on the scale of the projects'.

no official page stated this clearly

OK

SOC 2 / ISO 27001

= SOC 2 Type 2 (Favourable)

Announced in a July 2025 company blog post. The trust portal (trust.together.ai) renders in JavaScript and was not read, so ISO 27001 is not confirmed here.

Together AI has successfully completed a SOC 2 Type 2 examination.

FAIL

HIPAA BAA

= no (standard terms bar medical data) (Unfavourable)

The self-serve terms prohibit sending medical information. A company blog post mentions HIPAA practices and BAAs 'with our partners', but no customer BAA offer was found.

You will not use the Services to transmit or provide to the Company any financial or medical information of any nature

OK

Self-host option

= yes (Enterprise Platform) (Favourable)

Together Enterprise Platform can be deployed in your own VPC (AWS, Azure, GCP, OCI) or on-premise. Announced in a 2024 blog post; pricing and current availability not stated.

the Together Enterprise Platform can be deployed within your VPC or on-prem infrastructure.

Source checks

6 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