I abused PostHog's wizard to get free Claude tokens

A write-up titled “I abused PostHog’s wizard to get free Claude tokens” describes using PostHog’s setup wizard as a path to free Claude access. The piece is published at techstackups.com and was submitted to Hacker News as item 49147696, where it had 2 points and 0 comments at capture time. The framing centers on PostHog’s onboarding flow and Claude token access rather than on a formal product launch or disclosed vendor change.

The claimed path is product mechanics, not a published API or SDK feature: PostHog’s setup wizard is an interactive onboarding surface that can provision or expose Claude-related capability during setup. The article presents that surface as enough to obtain free Claude tokens when driven outside normal paid or quota-bound usage. No architecture diagram, benchmark, version pin, or exact token count is given in the available summary, so the technical claim stays at the level of wizard-mediated access rather than a measured exploit or performance result.

For engineers and builders, the signal is about free-tier and wizard design. Onboarding tools that spin up AI help, sample projects, or agent features can become a side channel for model usage if token budgets, account binding, and rate limits are not tied tightly to a verified, billable identity. Anyone shipping setup wizards that call Claude (or similar models) has to treat those steps as production cost surfaces, not just UX chrome.

In market context, PostHog sits in product analytics and tooling, while Claude is Anthropic’s model stack; free access through a third-party wizard sits at that intersection. The HN thread had almost no engagement (2 points, no comments), so this is an early, low-visibility claim rather than a widely validated incident. Related tooling names in the same feed cluster (Claude, Codex, Fable) only mark the broader agent/codegen news lane; they do not establish a multi-vendor campaign or coordinated disclosure.

What to watch next is whether PostHog tightens wizard-side Claude provisioning—auth, per-account caps, or removal of free model calls from setup—and whether follow-up posts or vendor notes confirm, deny, or patch the path. Until that lands, treat free Claude access via a product wizard as an unverified cost and abuse-risk story: verify against official PostHog docs and billing limits before relying on it, and if you run similar wizards, audit which steps call paid models and how those calls are scoped per user.