Canva — 137 Million Accounts, and Why "Passwords Exposed" Overstates It (2019)
If you have an account with Canva, here’s what’s now in circulation.
Canva hashed with bcrypt, and many users had no password at all because they signed in with Google or Facebook. This is the breach our own test suite cites as the example of a page that wrongly said "plaintext password leaked".
What happened
In May 2019 the graphic design platform Canva was breached, affecting 137,272,116 subscribers. The exposed data comprised email addresses, usernames, names, cities of residence and — for users not signing in through a social login — passwords stored as bcrypt hashes. The seller GnosticPlayers claimed responsibility and listed the data among the 2019 Dream Market batches.
The label that was wrong, and why we know
The HIBP catalogue records a data class called Passwords for this breach. That class means a password field was present in the dump. It says nothing about how the password was stored.
GalaxyWarden got this wrong once, publicly. Our own regression suite (tests/test_breach_record_count_is_derived.py) records the defect: a page rendered "Plaintext password leaked · Canva" from that data class. Canva's passwords were bcrypt. The label was false, and it was false in the direction that causes real harm — it tells a reader their password is public when it is not, which pushes them toward panic and away from the actions that would actually help.
There is a second layer here that automated pages routinely miss: a large share of Canva users had no password in the dataset at all, because they authenticated through Google or Facebook. For those accounts there was never a Canva password to leak, and "change your Canva password" is advice about something that does not exist.
What was actually at risk
Strip out the password panic and a real exposure remains: name, username, email address and city of residence, for 137 million people, many of them running small businesses, freelance design practices or social-media presences under a brand name.
City-level location paired with a real name and a working email address is a solid foundation for targeted fraud — particularly invoice and client-impersonation scams aimed at freelancers, who expect unsolicited mail from strangers and are professionally obliged to open it.
The identity-chain implication
Design-tool accounts sit at an awkward junction: registered under a real name for billing, used to produce work published under a brand or handle. The Canva dataset therefore links a legal name and a home city to a creative identity that may be deliberately separate from it.
For a creator whose brand is public but whose location is not, that is the entire distinction collapsing in one row — and no password change repairs it, because no password was involved.
What to do now
What You Should Do
- Check whether you used a social login — if so there was no Canva password, and the reused-password question does not apply to you
- If you did set a password, change it only where reused; bcrypt makes bulk recovery impractical
- Freelancers and small businesses: treat unexpected client or invoice mail as suspect, since name plus city plus email is exactly the fraud pretext this data supports
- Check the same address against Dubsmash, MyFitnessPal and MyHeritage from the same 2019 listings
- Review whether your billing name is linked to a brand identity you intended to keep separate
What the free scan actually returns
Found on people-search siteswe remove these
These listings are live, public, and legal to remove. That’s what a Deep Sweep buys.
Found in breach recordsverifiedreported — unverified
Each record is labeled: confirmed breach data, or an attacker’s claim no one has verified.
Leaked data cannot be deleted from the internet — anyone claiming otherwise is lying. Broker listings can be removed. We do the second, and show you exactly what to fix from the first.
Report details & sourcing
Related breaches
Zynga — 173 Million Words With Friends Accounts (2019)
The maker of Words With Friends lost 173 million accounts with salted SHA-1 passwords and, unusually…
MySpace — 360 Million Accounts, and the Weakest Password Storage of Any Major Breach (2008)
MySpace stored the SHA-1 hash of only the first ten characters of your password, lowercased, with no…
LinkedIn — 164 Million Accounts, Unsalted SHA-1, Four Years in the Dark (2012)
Hacked in 2012, sold in 2016. LinkedIn stored passwords as unsalted SHA-1 and the vast majority were…