Detailed Analysis
A user-reported account verification bug affecting Anthropic's Claude platform has created a frustrating authentication loop for at least some paid subscribers, preventing access to the desktop application and browser-based Claude interface even while Claude Code continues to function on the same account. The specific failure mode described involves an email verification link that redirects users to an account creation screen — one that simultaneously confirms the email is already verified — producing a logical contradiction that no amount of standard troubleshooting resolves. This category of bug is particularly damaging because it targets the authentication layer itself, meaning the user is entirely blocked from the product despite holding an active paid subscription and despite the underlying account data appearing intact on Anthropic's backend.
The verification issues described are not isolated incidents. Research context indicates that Anthropic's Claude platform has exhibited a broader class of account verification failures over an extended period, particularly around phone number verification during signup and login. A known policy restricts each phone number to a single Claude account, but the system frequently fails to communicate this restriction clearly, instead surfacing generic errors like "Unable to send verification code to this number" or "Error sending code. Double check your phone number." These opaque messages leave users without actionable guidance, and the absence of a dedicated ticketing or escalation pathway for authentication bugs forces affected users to file support requests in unrelated repositories — such as the Claude Code GitHub issue tracker — or navigate an AI-powered support agent that, by multiple accounts, cycles users through unhelpful scripted responses without resolution.
The broader significance of these persistent verification failures lies in the friction they introduce at the most critical point of the user funnel: account access. For a paid subscriber already committed to the platform, being locked out while Claude Code continues to operate on the same credentials underscores a fragmentation in how Anthropic's authentication systems handle different product surfaces. Claude Code appears to maintain session tokens or API-level access independently of the browser/desktop login flow, which means the verification bug is likely localized to a specific authentication pathway rather than a wholesale account deactivation. However, this architectural nuance provides little comfort to users who cannot access the primary interfaces they pay for.
From a competitive standpoint, authentication and onboarding failures carry disproportionate reputational weight in the AI assistant market, where OpenAI, Google, and others offer relatively frictionless account management. Community threads on Reddit and Hacker News reveal that some users have abandoned Claude entirely in favor of alternatives after encountering these verification barriers with no clear resolution timeline. The fact that these issues have been documented across GitHub issues, Reddit, and tech news aggregators for an extended period — without an apparent systematic fix — suggests either a deprioritization of authentication infrastructure work or deeper architectural complexity in how Anthropic manages identity across its growing product suite, which now spans Claude.ai, Claude Code, and API access tiers.
The situation also highlights a structural gap in Anthropic's support infrastructure relative to the scale at which it now operates. With Claude Code reaching a significant developer user base and Claude.ai serving a broad consumer audience, the lack of a dedicated, human-staffed authentication support channel creates compounding failures: users cannot self-serve their way out of verification loops, AI support agents cannot escalate appropriately, and there is no clear status communication about known bugs. Temporary workarounds — such as incognito browsing, alternative phone numbers, or simply waiting for server-side resolution — place the remediation burden entirely on the user, which is an untenable experience model for a paid product competing at the enterprise and prosumer level.
Read original article →