ChatPRD's Claire Vo says token fraud forced her to end free trials
Claire Vo says abuse of the ChatPRD free trial, often aimed at reselling tokens, has led her to wind down free trials entirely.

Watch the short
Under a minute, read by an AI anchor. Made .
YouTube loads only when you press play.
Claire Vo, who runs the AI product platform ChatPRD, says token fraud has become serious enough that she is winding down free trials entirely. In a YouTube video she described a flood of abuse from groups working the product's free trial. Some just wanted free AI usage. Others resold tokens or ran other commercial operations on top of the ChatPRD agent.
Vo's own post states what much of this abuse is after. She wrote that fraudsters are mostly after Opus tokens, obtained through resold promo codes or in the gap between a trial start and ChatPRD's fraud alerts.
Vo's post gives the scale in her own terms. She wrote that she spends 90% of her time on ChatPRD dealing with increasingly creative international token fraud operations, and that the fraud had cost her tens of thousands of dollars. She named an Indonesian fraud ring, an Egyptian code farm, and Chinese signups that had been on an automatic loop for the last 24 hours. She also said the company is profitable. The post does not give a dollar figure for total monthly losses, nor does it say how many accounts were involved.
Her account of her defenses shows why the problem is hard. She said she runs the firewall herself, uses Stripe Radar, has rate limits and many automated triggers, and has been through penetration testing and SOC 2 certification. The difficulty, in her telling, is that each control also catches legitimate users. Tightening one means causing annoyance or pain for real customers.
Vo's post says she has "some thinking to do." For now, free trials are "dead" on ChatPRD, she wrote. She said she might ban all non-US business in 2027 and may move toward 100% enterprise customers, whom she described as willing to pay more and as having more interesting product needs.
The pelican detail comes from Simon Willison, who wrote that one common prompt in the traffic asked for SVGs of pelicans riding bicycles. He suggested the prompt may be used to identify the underlying model, since different model families have a definite signature. He said he has used that idea before to guess the identity of stealth preview models on platforms such as OpenRouter.
What changed
ChatPRD is winding down free trials entirely after token fraud, with Claire Vo saying free trials are dead on the platform.
Type: Pricing and limits 路 Applies to: ChatPRD
Why it matters
Companies offering free AI trials or promo codes can have their compute resold or used for unrelated commercial work, which raises costs for the provider. Prompt patterns may also be used to identify which underlying models are being served.
What you can do
Teams that run free AI tiers should review how they limit abuse, such as usage caps and signup checks, before opening them to the public.
Original source
Where we saw it
Simon Willison
AI wrote this summary from the sources listed, and nobody reviewed it individually before publication. Check the original source before you rely on it. How it works