I asked Claude AI to collect Google Maps data for me. Ended up with a $1,310 bill in 48 hours.
A few months ago I had a simple idea: collect a database of coffee shops in Mexico — names, phone numbers, addresses. Nothing fancy. A personal project. I asked Claude (Anthropic's AI assistant) to help me build a script using the Google Maps Places API. Claude wrote the code, explained which API keys I needed, and told me not to worry about costs — Google gives $200-300/month in free credits,…
In a recent incident, an individual found themselves facing a staggering $1,310 bill after utilizing Claude, an AI assistant from Anthropic, to collect data from Google Maps. The person had initially sought Claude's assistance in building a script using the Google Maps Places API to gather information on coffee shops in Mexico, such as their names, phone numbers, and addresses.
Claude provided the necessary code, API keys, and reassured the person that Google provided ample free credits ($200-300/month), suggesting the project would comfortably remain within this limit.
Upon initiating the script, the person was soon met with a series of payment notifications. Initially, one small charge was declined due to insufficient funds, but subsequent payments amounted to a total of $1,310.05, incurred within a mere 48 hours. The person found this situation baffling, believing it to be a glitch in the system. Later, a final notification confirmed the $1,310.05 charge.
The cost of manually collecting the same data would have been considerably lower, amounting to $50-100, and would have also resulted in higher-quality data. The person's experience highlighted several key issues with Google Maps Platform. Firstly, there is no default spending cap, unlike other Google APIs such as Google Gemini, which have clear quota limits.
This lack of a cap means users can be charged unlimited amounts without any prior warnings, leading to unexpected, hefty bills for users like the individual in this story.
Furthermore, Google Maps Platform does not offer a free tier safety net, unlike some other Google APIs. There are no built-in protections to prevent accidental overcharges, leaving users vulnerable to such financial surprises. The person reached out to Google support, detailing the incident over weeks of emails and interactions with various agents and escalation managers.
Google acknowledged that the incident was accidental and approved a 75% refund of $982.53. However, $327.52 remained unpaid, and Google declared the case closed.
Despite the multiple attempts to escalate the issue through different channels, including email, Twitter, LinkedIn, and the Google Cloud Community, the person's efforts were largely futile. They received either automated responses or were redirected back to the billing team. This incident serves as a cautionary tale for independent developers and hobbyists who might rely on AI assistants like Claude to estimate costs for cloud API usage.
The lesson is clear: never trust an AI assistant's cost estimates for cloud APIs. Always verify the pricing information directly on Google's own website and ensure that hard budget limits are set before running any automated scripts.
The person also emphasized that Google Maps Platform poses significant risks for independent developers and hobbyists due to its lack of default caps and real-time warnings. An uncontrolled script could result in unexpectedly high bills, as evidenced by their own case. This unfortunate experience, cataloged as case #73959216, serves as a stark reminder for anyone considering using Google Maps Platform for their projects.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.