Anthropic's Supported Regions Policy Takes Effect Nov 12: Three-Layer Tightening Explained, and Why Celedog.io's Global Multi-Node Operation Is Unaffected

Celedog.io is served from 9 global nodes across Singapore, Germany, Australia, Indonesia and more, under a globalized operating entity — not a capital entity from an unsupported region — so global users can use it with confidence.

Policy in Effect: A Three-Layer Tightening of Claude's Availability

Anthropic's Supported Regions Policy took effect on November 12, 2026, tightening Claude's eligible regions and eligible users with clearer boundaries. Rather than mainly checking the access source, the new policy splits the determination into three layers — stricter coverage, clearer edges.

Core Change: Three Test Layers

  • Where the person is: It cannot be used while physically located in an unsupported region, including employees of supported-region companies on business travel to an unsupported region. The anchor is actual physical location, not account ownership.
  • Where the company is registered / headquartered: Entities registered or headquartered in an unsupported region, and their employees, cannot use it — even if the person is currently in a supported region.
  • Ownership and control: Entities directly or indirectly majority-owned or controlled by an individual or entity in an unsupported region also cannot use it.

Direct Impact on Users in China

For users in China the impact is fairly direct: mainland China, Hong Kong and Macau are not on the supported list; Taiwan is. The policy's coverage is broad, spanning Claude.ai, Claude Code, Cowork, the API and the developer platform, and it also covers Claude models delivered through third-party cloud platforms and authorized resellers — meaning "routing through a third party" does not automatically grant an exemption.

One distinction is critical: earlier, Claude's payment page briefly showed a mainland China option and accepted UnionPay cards, but this did not mean the service was opened. Eligibility still rests on the supported-regions list, account risk controls, and actual place of use; a transient payment-page display is not any commitment of availability.

Three Things Cross-Border Teams Should Verify

If you are building cross-border AI products, API aggregation, or multi-country sites, verify the following against the three layers of the new policy:

  • Actual access location: whether the user's real access IP and physical location fall in a supported region — travel and roaming count too.
  • Calling entity's domicile: whether the registered location and headquarters of the entity behind the calling account sit in a supported region.
  • Ownership and control: whether the underlying cap table and actual controllers involve entities in unsupported regions, including indirect holdings and control chains.

Have legal or compliance review the call, especially for overseas subsidiaries, third-party relay APIs, and proxied access — these are where the risk points cluster. Reducing "can I use it" to "can I connect" is the most common misjudgment under this policy.

Celedog.io: Globalized Multi-Node Operation — Use the Global Service with Confidence

After the tightening, the top question for developers became: will teams still calling mainstream models through a gateway get stuck because of the operating entity or node location?

As a developer-focused AI gateway and routing platform, Celedog.io is served from 9 global nodes including Singapore, Germany, Australia, and Indonesia, with a footprint spanning major regions across Asia, Europe, and Oceania for nearby, stable, auditable access. Celedog.io's operating entity has a globalized structure and is not operated by a capital entity from an unsupported region — on entity domicile, node distribution, and compliant access, it aligns with each upstream provider's policy requirements in a sustainable, auditable way. Therefore, global users can use Celedog.io's global services with confidence, without business interruption caused by policy shifts in any single region.

At the same time, individual developers and global teams don't need to handle per-vendor policy checks, entity review, and compliance communication on their own. Through a single OpenAI-compatible endpoint, they can keep mainstream models stably wired into their products, while Celedog.io's one-click deployment and massive Agent Skill library let the platform layer absorb model routing, multi-node scheduling, cost control, and failover — keeping the focus on product logic and business value.

One caveat: a globalized architecture and compliant access make the channel itself sustainable, but they do not replace your own entity-level compliance duty. Cross-border teams should still complete the three entity-and-location checks above and manage their own keys and data properly.

Conclusion

At its core, this Supported Regions Policy is Anthropic upgrading the availability test from "access source" to a three-layer "person + company + ownership" look-through. For individuals the boundary is clearer; for cross-border teams the compliance center of gravity shifts from "can I connect" to "do the entity and its controllers hold up."

For developers, rather than re-checking policy edges across a tangle of relays, proxies, and gray channels, place the gateway on a platform with a globalized operating entity, a compliant node footprint, and a paved compliant-access path. With 9 global nodes and a globalized operating entity, Celedog.io has reserved a stable lane for global users — every tightening only reaffirms the long-term value of a compliant base.


Last updated November 11, 2026

Where to go next