OPENAI

ChatGPT Access Check

See your network the way OpenAI does: is the exit chatgpt.com and the API actually see in a supported region, is it a VPN, proxy, datacenter or WARP, do all domains share one exit, and is your real IP leaking.

Check your ChatGPT access environment

We check, from OpenAI's side: whether the exit IP and region seen by chatgpt.com, api.openai.com and the sign-in domain are supported, whether that IP is flagged as a VPN, proxy or datacenter, whether your real IP leaks through WebRTC or DNS, and whether your system time zone contradicts the exit. Takes about 5–10 seconds.

Exit IP
Supported regions
VPN detection & IP leaks
Local network
Local system

No sign-in. IP intelligence and leak probes need network requests; browser settings are read locally

ChatGPT region and IP requirements at a glance

Whether ChatGPT works for you comes down to three things: first, is your exit IP in one of OpenAI's 208 officially supported countries and territories (mainland China, Hong Kong, Macau and Russia are not); second, is that IP flagged as a VPN, proxy, datacenter or Cloudflare WARP; third, do chatgpt.com, the API and sign-in use the same exit, and does WebRTC leak your real IP. Phone number, payment method and account history are account-level and can't be checked from outside.

OpenAI officially supports 208 countries and territories, and ChatGPT and the API share the same list (checked on 2026-10-21).
Not supported: mainland China, Hong Kong, Macau, Russia, Belarus, Iran, North Korea, Syria, Cuba and Venezuela. Taiwan, Singapore, Japan, South Korea and the United States are supported.
OpenAI states that accessing or offering access outside the listed countries may get your account blocked or suspended.
chatgpt.com, api.openai.com and auth.openai.com all sit behind Cloudflare; their traces show the IP and country OpenAI really sees.
ChatGPT's web app and mobile apps are more sensitive to datacenter IPs; running Codex CLI or the API on a cloud server is common, and the region rules are the same.
A DNS leak doesn't change the address OpenAI sees (queries only reach resolvers); a WebRTC leak hands the real IP behind your proxy to any page script.

Official lists checked on 2026-10-21OpenAI supported countries

Which countries and regions do ChatGPT / OpenAI support?

Official list

Based on OpenAI's official supported-countries page (ChatGPT and the API share one list): 208 countries and territories. Mainland China, Hong Kong, Macau, Russia, Belarus, Iran, North Korea, Syria, Cuba and Venezuela are not on it.

BelarusBY

ChinaCN

CubaCU

Hong KongHK

IranIR

North KoreaKP

MacaoMO

RussiaRU

SyriaSY

VenezuelaVE

KosovoXK

Anything not on the list is unsupported. OpenAI names French, Dutch and Danish territories one by one but not US territories such as Puerto Rico or Guam, nor UK territories such as Gibraltar; this page marks those as "Not listed". Ukraine is listed "with certain exceptions" (occupied areas are excluded), and IP location only resolves to the country, so this page can't tell them apart. The official list changes; the official page is authoritative.

How the score works

The score only covers what OpenAI can see from this visit, out of 100. An unsupported region is reported as "Unsupported region"; if chatgpt.com is unreachable the score is capped at 30; a Cloudflare vs. database disagreement where one side is unsupported caps it at 45; an unlisted territory caps it at 70. Everything else is weighted as below; unmeasured items are left out of the total, and below 85% coverage the verdict is never "Low risk".

Dimension	What we check	Why it matters	Weight
Exit network type	Whether the exit is home broadband, mobile, datacenter, VPN, proxy, Tor or WARP	Shared exits and anonymizing networks are the first thing risk systems look at	35
Real IP exposure	Whether the public address WebRTC reveals matches the exit ChatGPT sees	A second exposed address shows the real location behind the proxy	20
IP reputation	Abuse reports and DNSBL blocklist entries	Reputation follows the address; records left by previous users count against you too	15
Exit consistency	Whether exits match across domains, IPv4 / IPv6 countries and location sources	One account hopping between IPs and countries is a classic account-security anomaly	15
Time zone consistency	Whether the system time zone conflicts with the exit IP's time zone	Time zone is a location clue any page script can read without permission	10
Automation & cookies	Whether the browser is automated and cookies work	Automation traits trigger Cloudflare-style CAPTCHAs directly	5

The weights summarize risk signals common across the industry; they are not OpenAI documentation. ChatGPT's web app and mobile apps are noticeably more sensitive to datacenter IPs, so a datacenter exit earns only 30% of "Exit network type" here (50% on the Claude page), and a clean datacenter IP lands at "Usable, keep an eye on it".

OpenAI region and usage policies behind this page

Supported regions are a hard prerequisite

OpenAI's official page lists the 208 countries and territories where ChatGPT and the API are available, and states that accessing or offering access outside them may lead to your account being blocked or suspended.

Mainland China, Hong Kong, Macau and Russia are not on the list; access from them is refused regardless of how clean the IP is.
The list names many overseas territories individually (Réunion, Greenland, the Cayman Islands…); don't assume an unlisted territory follows its parent country.
This page prefers Cloudflare's location (ChatGPT sits behind Cloudflare) and cross-checks it against our own database.

Official supported countries

Shared exits and anonymizing networks raise verification odds

OpenAI hasn't published its risk rules. In industry practice, shared VPNs, proxy pools, Tor and frequent country hopping all raise the odds of CAPTCHAs and account reviews; ChatGPT's common "Unusual activity" notices and Cloudflare challenges are mostly about the exit.

That doesn't mean a VPN always gets you banned: what matters is whether the exit is shared by many users, flagged by commercial databases, and stable.
The same account showing up on different exits across the web app, mobile apps, API and sign-in is an inconsistency the platform can see directly.
This page's VPN / proxy / Tor, WARP and exit-consistency checks map to exactly these common signals.

The web app and the API tolerate datacenter IPs differently

Among IPs in supported regions, sign-up and sign-in on ChatGPT's web app and mobile apps are more sensitive to datacenter (IDC) IPs, while running Codex CLI or API requests from a cloud server is common; the region rules are the same.

Use home broadband or mobile data to sign up, sign in and add a payment method.
When Codex CLI or the API runs on a server, check that machine's exit (commands in the "Local system" card).
"Residential proxy" isn't home broadband: pooled, rotating residential exits get identified too.

Account-level risk controls are out of this page's reach

Your phone number's country, payment card billing address, sign-up country and usage history all live inside OpenAI's systems and can't be observed from outside.

Signing in for long periods from somewhere other than where you signed up is itself a risk signal.
A clean environment only means the network layer has no obvious problem; it doesn't mean the account is safe.
If your account is suspended or restricted, use the appeal channels in OpenAI's help center.

Usage policy violations also lead to bans

OpenAI's usage policies list prohibited uses; violations can lead to warnings, restrictions or termination of access, regardless of your network.

Bulk sign-ups, reselling accounts and offering access to people in unsupported regions are all high-risk.
This page only answers whether the network environment this visit exposes has problems.

Official usage policies

ChatGPT region & IP FAQ

Can I use ChatGPT in mainland China or Hong Kong?

Not directly. Mainland China, Hong Kong and Macau are not on OpenAI's official list (208 supported countries and territories); access from IPs there is refused and accounts may be suspended. Taiwan, Singapore, Japan, South Korea and the United States are supported.

Does OpenAI decide my region from my IP?

At access time, region is mainly decided by where your exit IP is located. chatgpt.com and the API sit behind Cloudflare, which passes the country it locates to OpenAI; the loc field in the trace this page reads is that value. At the account level OpenAI also looks at sign-up details, phone number and payment method, which can't be seen from outside.

Is an "Unusual activity" notice or repeated CAPTCHA an IP problem?

Usually it's about the exit: shared VPNs, proxy pools, datacenter IPs, Cloudflare WARP, or the web app and API of the same account using different exits all raise verification odds. Start with this page's "Exit IP profile" and "Local network" cards; if both look fine, check browser extensions, cookies and the account itself.

Do Codex CLI and the ChatGPT web app have the same IP requirements?

The region rules are the same: one shared list. The difference is IP type: running Codex CLI or the API on a cloud server is common, while sign-up and sign-in on ChatGPT's web app and mobile apps are more sensitive to datacenter IPs. Note that Codex CLI uses the exit of the machine it runs on; check that machine with the curl command on this page.

Will using a VPN with ChatGPT always get me banned?

Not necessarily. OpenAI hasn't published its risk rules, and the problem usually isn't the VPN itself but whether the exit is shared by many users, flagged by commercial databases or hops between countries. A dedicated, fixed exit in a supported region is far lower risk, and whatever you use, the exit must be in a supported region.

If the check says low risk, is my account safe?

No. This page only checks the network and browser environment this visit exposes; phone number, billing address, account history and usage live inside OpenAI and no outside tool can see them. A clean environment only rules out network-layer problems.

Why does this site see a different IP from OpenAI?

Your proxy app sends different sites through different exits (split routing). This page goes by chatgpt.com's own trace and scores the IP OpenAI sees; split routing itself costs no points, only different exits across OpenAI's own domains do.

Does this check log my IP?

IP intelligence lookups go through our server and create ordinary access logs; we don't sell or share lookup records. Browser details such as time zone, language and automation flags are read locally on your device and never uploaded.