Anthropic Adds ID Verification to Claude, Blocking Chinese Users

What Happened

Anthropic published a post titled "Identity Verification on Claude" announcing that a subset of users will now be required to submit government-issued ID documents and complete a real-time selfie for facial comparison before accessing certain Claude features, according to a report published on Ju ejin (掘金) citing the Anthropic announcement. The verification is conducted by an unnamed third-party company. Anthropic stated the measure is intended to "prevent model abuse and fulfill legal obligations."

Chinese mainland has appeared on Claude's list of unsupported regions since 2024, according to the same report. Because Chinese national ID holders are effectively excluded from the supported verification pool, triggering the verification flow produces one of two outcomes for affected users: failed verification resulting in restricted functionality, or — in cases where verification somehow proceeds — account termination.

The Juejin report frames this as the final step in a documented escalation spanning roughly two years:

• 2024: Claude terms of service first explicitly listed Chinese mainland as an unsupported region.
• February 2025: Accounts accessing Claude via OAuth- connected third-party tools were banned in bulk.
• H2 2025: Anthropic banned 1.45 million accounts; 52,000 appeals were filed, of which only 1,700 were overturned — an appeal success rate of approximately 3.3%.
• April 2026: Real-name ID and facial verification formally launched.

Why It Matters

The practical effect is a near-complete cutoff of Claude access for developers based in mainland China. The author specifically flags Claude Code — used heavily for day-to-day coding, debugging, and code review — as a tool that affected developers will need to replace.

The API- layer impact may be broader than consumer access. Projects and services built on the Claude API face service interruption if API credentials become tied to verified identities that Chinese-based operators cannot supply . Any product with Claude embedded as an inference backend is now carrying single-vendor geographic risk.

The pattern mirrors, in the author's framing, the U.S. chip export restrictions that accelerated domestic Chinese semiconductor investment. The comparison is not a market forecast but a rhetorical point: forced exclusion from a dependency tends to accelerate domestic alternatives. Domestic Chinese models — including Tongyi (通义), Doubao (豆包), and others — are cited as active iterators that stand to absorb displaced Claude users.

From a compliance standpoint, Anthropic's position is defensible given the current U.S.-China technology competition landscape. The company has not publicly characterized the move as geopolitically motivated; the stated rationale is abuse prevention and legal compliance. However, the cumulative architecture of the restrictions — territorial exclusion, OAuth ban, mass account purge, biometric gate — is struct urally indistinguishable from a deliberate market exit.

The Technical Detail

The verification flow requires two data inputs: a static government ID document and a real-time selfie matched against that document. Matching is handled by a third-party identity verification provider, not Anthropic's own systems, according to the Juejin report. The specific provider is not named in the available source material.

Trigg ering conditions — i.e., which user actions or account states invoke the verification gate — are not fully disclosed in the source article. This ambiguity is operationally significant: developers relying on Claude API keys issued under existing accounts do not yet have a clear picture of when or whether their credentials will hit the verification wall.

The OAuth-based workaround path , previously used to access Claude through third-party tool integrations, was closed in the February 2025 batch ban, per the report. That leaves direct API access and the Claude.ai web interface as the remaining vectors — both of which are now subject to the identity gate.

What To Watch

Over the next 30 days, the following developments warrant monitoring:

• Anthropic API key policy update: Whether existing Claude API keys held by users in restricted regions will require re-verification or face automatic revocation. No timeline has been disclosed in available sources.
• Third -party verification provider disclosure: Identity of the biometric vendor matters for enterprise data compliance assessments. Watch for Anthropic's privacy documentation update.
• Domestic Chinese model benchmark releases: Tongyi, Doubao, K imi, and DeepSeek are the most likely beneficiaries of displaced Claude Code users. Capability gaps in agentic coding tasks — Claude Code's core use case — will determine how quickly substitution occurs.
• Enterprise contract implications: Companies outside China that employ engineers in mainland China and provide Claude access as a workplace tool face immediate access management questions. HR and IT policy updates at multinational tech firms may follow.
• Appeals and community response: Given the 3.3% appeal overturn rate from the H2 2025 ban wave, large-scale organized appeals are unlikely to be effective , but watch for coordinated developer community responses on platforms like GitHub and Hacker News.