OpenCode Console breach exposes some users' AI provider keys
Hackers reached part of the OpenCode platform and stole API keys for a small share of its users. Revocation notices have gone out to affected accounts.
By Standard Topic. AI-generated from the reporting listed under Coverage. Published . Updated .
This story is no longer being updated. It reflects reporting up to .
What happened
A breach of the OpenCode Console, the online platform for the open-source AI coding agent OpenCode, may have exposed API keys for OpenAI, Anthropic and Google services. Hackers also ran up charges on some users' AI accounts.
Why it matters
Exposed keys can be used to run up charges on users' AI accounts, so affected users need to revoke them at the provider.
Timeline
- Mid-October 2026: A breach of the OpenCode Console exposes private API keys for OpenAI, Anthropic and Google services of about 1.4% of Console users.
- Oct 19, 2:07 AM ET: OpenCode says the breach may have exposed AI provider keys and sends revocation notices to affected users.
- Oct 19, 3:28 PM ET: Hackers are reported to have run up charges on users' AI accounts using the stolen keys.
Coverage
- Hackers hit AI coding tool OpenCode, running up charges on users’ AI accounts (techcrunch.com)
- OpenCode Breach Could Affect OpenAI, Anthropic and Google (Kevin Poireault)
- OpenCode Console Breach May Have Exposed OpenAI, Anthropic, and Google API Keys (thehackernews.com)
- OpenCode Console Breach Exposes OpenAI, Anthropic, and Google API Keys in Credential Theft and Distillation Attack (October 2026) (rescana.com)
- OpenCode Console Breach Affects 1.4% of Users, May Have Exposed OpenAI and Claude API Keys (xenospectrum.com)
- OpenCode Security Breach Exposes third-party API keys for its Console users (cybersecuritynews.com)
- OpenCode breach exposes stored OpenAI and Anthropic keys, prompts urgent key rotation (startupfortune.com)
More from Technology & AI
- Eight states sue OpenAI, alleging misleading AI safety claims
- OpenAI announces solution to Navier-Stokes math problem
- Datadog bets AI agents can handle testing of AI-written code
- Anthropic and OpenAI sketch wealth-sharing plans before IPOs
- Mamdani administration proposes AI customer service rules
- Florida engineers study how AI behaves in high-risk situations