AI Agent Allowlist
Home Page-Types Database Agent Guardrails 2026 Incidents API Docs Pricing
Resources
Use Cases (15) Industries & Buyers (12) Learn: Core Concepts (12) Implementation Guides (15) Comparisons (8) Agent Security Guides (22) Market & Frameworks (9) Schema & Data Reference (6) FAQ Glossary
Why It Matters
2026 Agent Incidents Category Targeting Database Refreshes Contact Customer Login
Download Free Sample
AI Agent Allow List: Agent Data-Use Register

Does this AI agent train on your data?

The AI agent allow list checked 1,413 agent and coding-tool vendors against their own published terms for whether the agent trains on what it sees, dated 4 November 2026.

A chatbot sees what you type. An agent can see a whole repository, a whole screen, or a whole logged-in session.

464do not train
161train by default
72train unless opted out
716not stated

40 vendors, checked

Coding agents and agent platforms first, then by domain. A vendor not in this table has not yet been checked in the published register.

Every value below is our own classification, derived from the vendor's own published privacy policy, terms of service or terms of use, not the vendor's wording. Search that vendor's own published terms to verify directly. See our accuracy note and correction path.

VendorCategoryTrains on dataOpt-outEnterpriseAPICheckedOn file
Cursor
cursor.com
Agents & AutomationNoYesYesNot stated17 Sep 2026 Checked against the vendor's own policy
pull requests
github.com
Agents & AutomationTrains unless opted outYesYesNot stated17 Sep 2026 Checked against the vendor's own policy
ChatGPT
openai.com
Code & DevelopmentTrains unless opted outYesYesYes17 Sep 2026 Checked against the vendor's own policy
Devin
devin.ai
Code & DevelopmentNot statedNot statedNoNot stated17 Sep 2026No source on file
Zapier Central
zapier.com
Agents & AutomationTrains unless opted outYesYesNot stated17 Sep 2026 Checked against the vendor's own policy
n8n
n8n.io
Agents & AutomationNoNot statedYesNot stated17 Sep 2026 Checked against the vendor's own policy
browser-use
browser-use.com
Agents & AutomationYesNot statedNot statedNot stated17 Sep 2026 Checked against the vendor's own policy
Bolt x Figma
bolt.new
Agents & AutomationNot statedNot statedYesNot stated18 Sep 2026No source on file
Grok API
x.ai
Agents & AutomationYesNot statedNot statedNot stated17 Sep 2026 Checked against the vendor's own policy
11.ai
11.ai
Agents & AutomationTrains unless opted outYesNot statedNot stated17 Sep 2026 Checked against the vendor's own policy
fx (by Vercel)
fx.sh
Agents & AutomationNot statedYesYesNot stated18 Sep 2026No source on file
Pi
pi.ai
Agents & AutomationTrains unless opted outYesNot statedNot stated18 Sep 2026 Checked against the vendor's own policy
OpenMarket
m11.ai
Agents & AutomationNot statedNot statedno_tierno_tier18 Sep 2026No source on file
Any.do
any.do
Agents & AutomationNot statedNot statedNot statedYes18 Sep 2026No source on file
omiGPT
omi.me
Agents & AutomationNoNot statedNot statedNot stated18 Sep 2026 Checked against the vendor's own policy
VDF AI Networks
vdf.ai
Agents & AutomationNot statedNot statedNoNot stated18 Sep 2026No source on file
Ito AI
ito.ai
Code & DevelopmentNoNot statedno_tierno_tier18 Sep 2026 Checked against the vendor's own policy
sim
sim.ai
Agents & AutomationNoNot statedYesYes17 Sep 2026 Checked against the vendor's own policy
Qvio™️
hia.ai
Agents & AutomationNot statedNot statedno_tierno_tier17 Sep 2026No source on file
Exa
exa.ai
Agents & AutomationYesNot statedYesYes17 Sep 2026 Checked against the vendor's own policy
11x.ai
11x.ai
Agents & AutomationNot statedNot statedYesNot stated18 Sep 2026No source on file
Furl
furl.ai
Agents & AutomationNoNot statedNot statedNot stated18 Sep 2026 Checked against the vendor's own policy
Aito
aito.ai
Agents & AutomationNoNot statedNoNot stated18 Sep 2026 Checked against the vendor's own policy
Gedo
gedo.ai
Agents & AutomationNoYesno_tierno_tier18 Sep 2026 Checked against the vendor's own policy
Plow Latch
plow.co
Agents & AutomationNoNot statedNot statedNot stated18 Sep 2026 Checked against the vendor's own policy
Cust
cust.co
Agents & AutomationNot statedNot statedYesno_tier18 Sep 2026No source on file
Auxx.ai
auxx.ai
Agents & AutomationNot statedNot statedYesYes18 Sep 2026No source on file
Aira
aira.io
Agents & AutomationYesNot statedno_tierno_tier18 Sep 2026 Checked against the vendor's own policy
Awaz.ai
awaz.ai
Agents & AutomationTrains unless opted outYesNoNot stated18 Sep 2026 Checked against the vendor's own policy
Qodo
qodo.ai
Code & DevelopmentTrains unless opted outYesYesYes17 Sep 2026 Checked against the vendor's own policy
Fimo
fimo.ai
Agents & AutomationNot statedNot statedYesNot stated18 Sep 2026No source on file
Felo
felo.ai
Agents & AutomationNot statedNot statedNoNot stated18 Sep 2026No source on file
Enjo.ai
enjo.ai
Agents & AutomationNoNot statedYesYes18 Sep 2026 Checked against the vendor's own policy
Effy
effy.ai
Agents & AutomationNoNot statedYesNot stated18 Sep 2026 Checked against the vendor's own policy
Kore.ai
kore.ai
Agents & AutomationNot statedNot statedYesNot stated18 Sep 2026No source on file
Oumi
oumi.ai
Code & DevelopmentNot statedNot statedYesNot stated18 Sep 2026No source on file
Gong.io
gong.io
Agents & AutomationNot statedNot statedYesNot stated18 Sep 2026No source on file
Gong.AI
gong.ai
Agents & AutomationNot statedNot statedYesNot stated18 Sep 2026No source on file
is.team
is.team
Agents & AutomationNoNot statedNot statedNot stated18 Sep 2026 Checked against the vendor's own policy
Kilo
kilo.ai
Code & DevelopmentYesYesNot statedNot stated17 Sep 2026 Checked against the vendor's own policy

Why agents see more than chat

The whole repository

A coding agent with repo access can read files a chat window never would: configuration, credentials left in code, internal comments.

The whole screen

A browser agent captures whatever is on screen during a session, including logged-in pages the user never intended to share.

Credentials in context

Agents often operate with the same session tokens and API keys as the user, widening what a training exception could actually expose.

What the register answers

Four fields per vendor, with values and a clause on click: trains on data, opt-out available, enterprise tier excludes training, API tier excludes training.

Seven more fields are locked and shown by name, unlocked in the full register.

What the allow list still controls

Regardless of what a vendor's terms say about training, the AI agent allow list keeps agents restricted to approved page types, so a training gap in one vendor's policy is not the only line of defense.

See the page-type database this runs on.

FAQ

What is the AI agent allow list register?

A dated record of what each AI agent and coding-tool vendor's own published terms say about training on what the agent sees: your code, your screen, your files.

Why does this matter more for agents than for chatbots?

A chatbot sees what you type. A coding agent can see a whole repository. A browser agent can see a whole logged-in session. The training question applies to a much larger surface.

How many agent vendors train by default?

161 of the vendors checked state they train on submitted data by default, 72 train unless the user opts out, 464 state they do not train, and 716 do not state a position, as of this page's snapshot.

What does the AI agent allow list do about this?

It keeps agents restricted to approved page types regardless of what any individual vendor's terms say, so a training gap in one vendor's policy is not the only control in place.

What are the locked fields?

Seven additional fields, including opt-out scope, feedback exceptions and monitoring retention, are shown by name only here and unlocked in the AI Tools Blocklist register product.

See every field, unlocked

The full register adds opt-out scope, retention, human review and subprocessor training for every vendor above.