Overview
- Users worldwide reported that signups, logins and chats failed after errors began around 8 PM ET on Wednesday, August 19.
- OpenAI’s status page lists as many as 12 affected API endpoints and shows that its Codex coding service and the chat web client are also impacted.
- The company acknowledged the incident publicly, said engineers identified the source of the problem, and said mitigation work is underway while the outage remains active.
- Real‑time monitors and search trends recorded sharp spikes in outage reports and diagnostic queries as thousands of users and developers ran into authentication and “too many concurrent requests” errors.
- Because ChatGPT and the OpenAI API power many consumer tools and developer workflows, prolonged disruption could stall coding projects and apps that rely on affected endpoints so users should follow the status page for restoration updates.