Apify says stolen credential exposed customer accounts and data
Original title: Apify Security Incident
Access on 9–10 November was cut off on 11 November; identified customers were notified and a post-mortem is due within a week.
On 9 and 10 November 2026, an unauthorized party used a compromised credential to obtain account credentials and access related Apify accounts. Apify detected the activity on 11 November at 12:13 UTC, closed that path, and says it has seen no further attacker activity since then. Customers whose personal information or stored data was accessed have been emailed. API tokens are now stored only as hashes. The incident affects the API, Actors, and general services.
Unconfirmed: The full scope of affected accounts and the root cause are still under investigation.
- Apify revoked some OAuth connections, including GitHub, as a precaution.
- Users who build Actors from private GitHub repos should reconnect GitHub.
- Open storage remains readable by anyone who knows the storage ID.
- A post-mortem with the root cause is promised within a week of the 13 November update.
Impact: Account credentials and some customer personal or stored data accessed
| Product | Affected |
|---|---|
| Apify | — |
- 1.Set General resource access to Restricted so storage IDs are not world-readable.
- 2.Reconnect GitHub if you build Actors from private repositories.
- 3.Follow Apify’s email if notified, and reconnect integrations if prompted.
- ATT&CK
- T1078 Valid Accounts
- Vendors
- Apify
From the original article
3 min read at status.apify.comUpdate: Our investigation and response are ongoing. What we have done since the last update - API tokens are now stored only as hashes, which limits the impact of any similar incident. - Added additional security layers to block the attack paths, including network restrictions between internal services and stronger authentication on internal systems.
By cynicalpeace · status.apify.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.