Skip to main content
  1. Home
  2. Computing
  3. News

A simple coding mistake is exposing API keys across thousands of websites

Security gaps that are easier to miss than you think

Add as a preferred source on Google
Computer, Electronics, Laptop
Adobe Stock Image

After analyzing 10 million webpages, researchers have found thousands of websites accidentally exposing sensitive API credentials, including keys linked to major services like Amazon Web Services, Stripe, and OpenAI.

This is a serious issue because APIs act as the backbone of the apps we use today. They allow websites to connect to services like payments, cloud storage, and AI tools, but they rely on digital keys to stay secure. Once exposed, API keys can allow anyone to interact with those services with malicious intent.

Sensitive API keys exposed across thousands of sites

According to TechXplore, the researchers identified 1,748 unique API credentials across nearly 10,000 webpages, tied to 14 major service providers. These leaks were not limited to obscure sites, with some appearing on platforms run by global banks and major software developers.

Recommended Videos

Around 84% of these leaks came from JavaScript files, which are easily accessible through a browser. This means the credentials were effectively sitting in publicly visible code.

Even more concerning is how long these keys remained exposed. Some were visible for up to 12 months, while a few rare cases showed credentials staying public for several years without detection.

So, what’s causing these leaks?

The study makes it clear that the problem does not lie with service providers like Amazon, Stripe, or OpenAI. Instead, the issue stems from how developers handle API keys.

In many cases, developers accidentally include private API credentials in the front-end code of a website, leaving it visible to anyone who knows where to look.

How to stop API keys from being exposed?

To prevent future leaks, the researchers suggest a few practical steps. Developers should scan the live version of their websites, and not just private code, to catch exposed keys.

With the rise of vibecoding, companies need stricter rules for automated website-building tools that handle sensitive data during deployment. This is also why platforms like Lovable have started adding safe browsing tools to protect users from poorly vibecoded websites.

Meanwhile, service providers need to improve detection systems to flag exposed keys the moment they appear online. Although responsible disclosure helped reduce some of these leaks, the scale of the issue remains significant.

Recent reports have also shown how simply visiting a website can expose your device to serious risks, highlighting how fragile web security can be for everyday internet users.

Manisha Priyadarshini
Manisha Priyadarshini is a tech and entertainment writer with over nine years of editorial experience.
Substack now lets you check if a post was written by AI
A new Pangram-powered scanner lets you check posts, notes, and replies for signs of AI writing.
Video playing on Substack.

Substack is giving readers a way to check whether the post they're reading was written by a human or by a chatbot. The company has partnered with AI-detection firm Pangram to introduce new tools that will let users scan posts, notes, and replies for AI-generated text. CEO Chris Best introduced the features in a post titled "Against Claudefishing," his term for content that leans on AI while presenting itself as human work.

How the scanning tool works

Read more
China’s AI talent shortage has tech giants recruiting teenagers
Forget campus recruiting, china's biggest tech firms are betting on teenage coders.
Artificial Intelligence

A 13-year-old boy in Hangzhou has already won national AI competitions and built a following of more than 136,000 people online, all while his dad tries to figure out how to guide him through a field that barely existed when he himself was growing up. That family's situation, first reported by Rest of World, says a lot about where China's tech industry is heading right now.

Companies used to wait for graduates to walk through the door. Now they're reaching further back, first to undergrads, and increasingly to teenagers, hoping to spot rare talent before anyone else gets to them.

Read more
OpenAI says AI models autonomously pulled off a major hack, but only a Chinese AI helped recovery
OpenAI

OpenAI’s latest cybersecurity test produced a result that sounds like a cautionary sci-fi script. Its AI models managed to escape their sandbox and reached the open internet. This is where things took a scary turn as it began hacking Hugging Face to steal the answers to the test they were taking.

The company says GPT-5.6 Sol and a more capable unreleased model autonomously chained together vulnerabilities across OpenAI’s research systems and Hugging Face’s production infrastructure. OpenAI has described the event as an unprecedented cyber incident.

Read more