Breach · · 3 min read
OpenCode Console breach: stolen AI API keys ran up users’ bills
Attackers got into an older OpenCode system and took AI provider keys that 1.4% of Console users had saved. Some people saw hundreds of dollars in charges overnight.
OpenCode, an open-source AI coding assistant, says attackers broke into an older part of its online platform overnight and took API keys that some users had saved there. Those keys were then used to run up charges on people’s OpenAI, Anthropic and Google accounts. If you use OpenCode Console, or keep AI keys in any online tool, here’s what to know.
What happened
OpenCode said on Saturday, October 17, that a breach the night before affected 1.4% of OpenCode Console users, people who had stored their own AI-service API keys in Console workspaces. It emailed everyone affected and said anyone who didn’t get an email wasn’t affected, TechCrunch reported.
Overnight, developers posted spending alerts and unexpected charges. Some reported hundreds of dollars on a single account; a few said more than $1,000. OpenCode described it as part of a “large-scale distillation attack,” meaning the keys may have been used to generate large volumes of AI output; some critics questioned that framing. Anthropic and Google blocked the keys they believed were exposed. OpenAI said it worked with OpenCode to restore access and refund its customers, and OpenCode says refunds for unauthorized charges are underway.
Co-founder Dax Raad said only the hosted Console was affected, not the open-source app or command-line tool, whose keys never reach OpenCode. OpenCode hasn’t said how attackers got into the older system or how the keys were stored, and said the incident was still ongoing.
What this means for you
An API key is a password that spends money. Unlike a stolen card, providers usually don’t flag unusual use, so a leaked key can burn through a budget in hours. And keys tend to live in many places at once: a coding tool, a config file, a cloud dashboard.
What you can do
- Revoke and replace any key you stored in OpenCode Console, directly in each provider’s dashboard, as OpenCode advises. Do it even if you aren’t sure.
- Set spending limits and alerts with every AI provider you use. Many users only noticed because their alerts went off.
- Use separate keys for separate tools, so one leak doesn’t expose everything and is easy to cut off.
- Check your provider bills for the past week and ask for a refund on anything you didn’t do.
- Keep an eye out for related phishing. “Your key was leaked, sign in here” emails often follow incidents like this.
How Frame helps
Frame focuses on personal data rather than developer keys, but the habit is the same: know which services hold what, and cut access the moment something leaks. Frame’s breach catalog lets you add incidents you were part of to a verified email and get a short checklist for each. Automatic breach lookups aren’t live yet. For more on how credentials get stolen in the first place, see how infostealers steal sessions.
Sources
- TechCrunch, Hackers hit AI coding tool OpenCode, running up charges on users’ AI accounts, October 19, 2026
New posts, by email.
About once a month. Plain-language privacy notes, nothing else. I never share your address.