Customers of a web-scraping platform with roughly 80,000 accounts, including Microsoft, Samsung, T-Mobile, Accenture, Siemens, the European Commission, and Roche, were told to reset 2FA after an internal intrusion. According to BleepingComputer, an unauthorized third party used a compromised credential on 9 and 10 November 2026 to reach parts of Apify’s internal systems, obtain additional account credentials, and access related Apify accounts. Passwords, Apify told affected users, remained safe. Second-factor codes may not have. Device-bound, phishing-proof authentication would have denied that transferable internal login and left no server-side OTP material to collect; rotating customer API tokens and sessions after detection is still separate containment once a foothold already exists.

What Apify confirmed after the November 2026 intrusion

Apify’s incident language is short and specific. According to the company’s update as reported by BleepingComputer, on 9 and 10 November 2026 an unauthorized third party gained access to parts of Apify’s internal systems. From that position the attacker obtained additional account credentials and reached related Apify accounts. The Apify status page and customer notices describe the same arc: internal access first, broader account reach next, containment after detection.

Detection hit at 12:13 UTC on 11 November 2026. Apify cut off attacker access, invalidated internal credentials, rotated customer API tokens, rotated Apify Proxy passwords, and reset active sessions. Some third-party OAuth connections, including GitHub, were revoked as a precaution. Investigation found OAuth was not the attack vector. No further attacker activity was reported after access was cut off.

Public reporting has not established how the first credential was compromised. AiTM phishing, a spoofed login page, helpdesk recovery abuse, password spraying, and credential stuffing are not proven paths in the sources used here. What is established is that a credential worked on an internal path, more credentials followed, and customers were advised to reset 2FA because codes may have been exposed. The full attack narrative and timeline sit on the companion post at legacymfa.sucks.

Why transferable internal secrets and reachable 2FA codes failed

The first failure was a working internal login secret in unauthorized hands. Whatever produced that secret, the login path still accepted a transferable credential. Once that path succeeded, the attacker could obtain additional account credentials from systems reachable after sign-in. That is the lateral step Apify described: related accounts opened because more recoverable secrets sat behind the first door.

The second failure is the one customers felt directly. Apify’s customer letter, as reported by BleepingComputer, said 2FA codes may have been exposed while assuring users that passwords remained safe. Password safety does not clear second-factor risk when OTP codes, seeds, or other recoverable factors live in systems an attacker already reached. Customers were told to reset 2FA. Individual notifications continued after the 12 November 2026 public reporting wave. One customer notice said there was no indication that account data stored on the platform had been affected. Quantified counts of hit accounts are not in public sources, and stored-data scope remained under investigation.

Stage What was at risk Transferable factor
Internal login Compromised credential used 9-10 Nov Secret that still worked for an outsider
Related accounts Additional credentials obtained Recoverable account secrets after foothold
Customer 2FA Codes may have been exposed Server-side or recoverable second factors
API and sessions Tokens and proxy passwords rotated Long-lived secrets after authentication

API tokens, Apify Proxy passwords, and live sessions sit after a successful authentication. Containment correctly rotated them and reset sessions. That work stops reuse. It does not erase the days the credential path was live. Shortening token lifetime after the fact would not have blocked the original internal login either. Public reporting does not name every MFA product on every Apify account path. It does show that 2FA codes were treated as potentially exposed artifacts, which is exactly what recoverable second-factor designs allow once internal systems are open.

Where device-bound credentials change the Apify path

Phishing-proof authentication built on device-bound public-key credentials changes the first two stages. There is no password or shared secret to replay on an internal login. The private key never leaves the enrolled hardware. The server holds only public keys. An unauthorized party who somehow obtains a string that used to be “the credential” still cannot complete a signature the IdP will accept, because nothing transferable exists to present.

The same design removes the 2FA-code exposure surface Apify had to warn about. There is no OTP to read out of a database, no SMS or email code sitting in a log, and no push approval an insider path can harvest. Second-factor material is not a server-side secret waiting for the next foothold. That is Prevention, Not Detection: the login never completes for someone who lacks the enrolled device, so the window that produced extra account credentials and reachable codes does not open.

Control Internal credential reuse Exposed 2FA codes Already-issued API tokens
Password plus OTP or push Attacker can present stolen secret Codes or seeds may be readable Revoke and rotate after the fact
Device-bound phishing-proof MFA No transferable secret to present No server-side OTP material Still requires token revoke

The third stage is different. Once an attacker already holds internal position and related account credentials, long-lived API tokens and live sessions are a control-plane problem. Device-bound login does not reach into memory, token stores, or proxy password tables after authentication has already succeeded. Apify’s rotation of customer API tokens, Apify Proxy passwords, session reset, and precautionary OAuth revocation were the right containment moves for that stage. Upstream value is clear: deny the transferable internal factor so the attacker never reaches the plane where those secrets live.

Workforce identity failures cascade into customer impact when internal credentials, account material, and API control planes share trust. Apify customers run scrapers, actors, and integrations that hang off those tokens. Blocking the internal credential path is therefore both an employee-login problem and a customer-token prevention problem. The architecture that keeps private keys on device is what keeps that cascade from starting.

Key takeaways for defenders

The next Apify-style window closes when internal login stops accepting anything an outsider can carry away, and when second factors stop existing as copyable codes. Token rotation remains mandatory cleanup for secrets that outlive a bad day. Authentication design decides whether that day starts.

FAQ

Would phishing-proof MFA 2.0 have stopped the Apify breach?

Phishing-proof MFA 2.0 would have blocked the transferable internal credential path and removed server-side 2FA code exposure in the Apify November 2026 incident. According to BleepingComputer’s reporting of Apify’s update, an unauthorized party used a compromised credential on 9 and 10 November 2026, then obtained additional account credentials. Device-bound keys leave nothing transferable to present at that login and no OTP material to harvest afterward. Already-issued API tokens and sessions still required rotation after detection, which no login MFA undoes.

Why did Apify tell customers to reset 2FA if passwords remained safe?

Apify told customers to reset 2FA because 2FA codes may have been exposed even though passwords remained safe. The customer letter summarized by BleepingComputer separated those two surfaces on purpose. Recoverable second-factor codes can sit in internal systems after a foothold, so password safety alone does not clear the second-factor risk customers had to address.

Does device-bound login prevent theft of Apify-style API tokens after authentication?

Device-bound login does not prevent theft or reuse of API tokens and live sessions after authentication has already succeeded in an Apify-style incident. Apify rotated customer API tokens and Apify Proxy passwords and reset sessions after detecting activity on 11 November 2026. That containment is required once long-lived secrets exist. The prevention work is earlier: stop the compromised-credential foothold so the attacker never reaches those control planes.

Are passkeys enough to cover the Apify failure modes?

Passkeys are origin-bound at login because they use origin-bound, hardware-held signatures, but they are not automatically phishing-proof across enrollment and recovery. Apify’s public record centers on a compromised internal credential and possible 2FA code exposure, not a named passkey deployment failure. If recoverable OTP material still exists in internal systems, passkeys at login alone do not remove the exposure Apify warned customers about. Full-lifecycle coverage means phishable factors never appear at registration, onboarding, or authentication.

Was OAuth the way attackers entered Apify?

OAuth was not the attack vector in the Apify breach. Apify revoked some third-party OAuth connections, including GitHub, as a precaution after detection on 11 November 2026. Investigation found OAuth was not how the unauthorized party gained access. The established path was a compromised credential into internal systems, followed by additional account credentials.