Zeabur security incident exposes stored API keys as users report unauthorized AI charges

Zeabur security incident exposes stored API keys as users report unauthorized AI charges

N
News Editor
2026-08-30 08:45:29
Taiwan-based cloud deployment platform Zeabur has disclosed a security incident involving unauthorized access to user project environment variables, raising the risk that stored API keys, access tokens, and third-party credentials were taken. The company said its monitoring system detected unauthorized use of an internal service credential on Aug. 27, which was then used to query data in its database. Zeabur stated that it currently has no evidence that account passwords, personal data, server logs, other project files, or credit card numbers were accessed. Details compiled from Zeabur’s status page and the security blog lonelysec indicate the attack path may have started with an AWS administrator credential, then moved through a shared cluster in Tokyo and a control-plane VPN before reaching the main database. According to lonelysec, the exposed credentials may include keys linked to OpenAI, Anthropic, OpenRouter, AWS, GitHub PAT, Cloudflare, Linode, DigitalOcean, Stripe, database passwords, and JWT secrets. Some users have already reported unexplained charges on AI service accounts. In response, Zeabur has suspended its AI Hub service after detecting suspicious LiteLLM-related activity and said a full post-incident report will be released later. Claims circulating on forums about a "612GB compressed customer database," source code, GCP owner credentials, and Google Workspace admin sessions remain unverified.

Zeabur, a Taiwan-based one-click deployment platform with more than 100,000 registered users, has disclosed a security incident that may have exposed customer API keys stored in project environment variables. According to Zeabur’s status page and reporting compiled by the security blog lonelysec, some users have already reported unexplained charges on OpenAI, Anthropic, and OpenRouter accounts.

Unauthorized use of an internal service credential

Zeabur said its monitoring system detected unauthorized use of an internal service credential on Aug. 27. The attacker used that credential to connect to a database and query project environment variables stored for users. The company said some of that data may have been taken, including user-defined API keys, access tokens, and third-party credentials.

On its status page, Zeabur classified the incident as "unauthorized access to project environment variable data." At the same time, the company said it has no evidence at this stage that Zeabur account passwords, personal information, server logs, other project files, or credit card numbers were accessed.

lonelysec traces the attack path back to AWS credentials

lonelysec, a security blog written by Bobo, outlined a possible attack chain that began with an AWS administrator credential. From there, the attacker allegedly moved through a shared cluster in Tokyo and a control-plane VPN, then reached the primary database and obtained customer environment variables.

According to lonelysec’s breakdown, the exposed credential set may span a wide range of services, including OpenAI, Anthropic, OpenRouter, AWS keys, GitHub personal access tokens, Cloudflare, Linode, DigitalOcean, Stripe, as well as database passwords and JWT secrets. That helps explain why some users saw abnormal billing on AI services: stolen API keys were apparently used to consume usage quotas on victim accounts.

AI Hub paused; forum claims remain unverified

Zeabur has suspended its AI Hub service after detecting suspicious LiteLLM-related activity. The company also said it will publish a full post-incident report, though that report had not been released at the time of the source article.

Separately, forum posts have surfaced offering what sellers described as a "612GB compressed customer database" and claiming access to source code, GCP owner credentials, and Google Workspace administrator sessions. Those claims remain unverified. Zeabur said it has no "direct evidence" of a large-scale data breach beyond Variables.

Users are being urged to rotate stored secrets

For users who stored credentials on Zeabur, the practical response described in the source is to act on the assumption that those secrets may already be exposed. That means revoking and regenerating API keys and tokens stored in environment variables, including those tied to OpenAI, Anthropic, OpenRouter, AWS, GitHub, and Stripe, then reviewing recent usage and billing records for anomalies and rotating database passwords and JWT secrets.

The incident also highlights a structural risk in one-click deployment workflows: when developers centralize a large number of high-privilege keys on a single platform for convenience, a breach of that platform can turn into a gateway to many other services at once.

This article was originally published by Bit.Fan. For more cryptocurrency news and market insights, visit www.bit.fan.
1200

Disclaimer:

The market information, project data, and third-party content displayed on this platform are for industry information sharing only and do not constitute any form of investment advice or return commitment.

Cryptocurrency trading carries high risks. Users should fully assess their risk tolerance and make independent decisions. All profits, losses, and legal responsibilities are borne by the users themselves.