TL;DR: A password manager has exactly one job — be there when you need to log in — and paying users report it silently failing that job. PainHunt's Security cluster describes broken Safari autofill and post-re-auth lockouts with no recovery path. The opportunity is not another vault; it is a reliability and recovery layer that sits beside whatever manager you already trust.
The evidence
Password Management is the highest-intensity cluster in PainHunt: average intensity 8.4/10 across 219 high-signal posts (average score 11.4/15), arriving mostly through Google Play (37) and the App Store (17), with a long tail on Mastodon and Hacker News. These are paying users writing at the moment the tool breaks.
The complaints split into two failure modes, both alarming for a security product. The first is the browser extension quietly dying: users describe a Safari extension as "literally unusable" despite an active subscription, autofill completely non-functional so credentials must be copied and pasted by hand, no interaction between the extension and the page being viewed, and no fix even on a clean OS install. The second is lockout: after a forced re-authentication, users cannot log back into a vault even with the correct password, the reset path does not work or is not offered, and there is no recovery route — a complete loss of access to the one place every other credential lives.
The requested features are strikingly modest. Not new capabilities — reliability: an extension that actually autofills, page interaction that works, a password reset that functions when standard login fails, and a real account-recovery process for locked-out users. People are not asking for more; they are asking for the thing they already bought to keep working.
Why now
Browser and OS platforms tightened the extension sandbox. Safari's extension model, macOS permission prompts, and repeated re-authentication requirements have made the surface a password manager depends on more fragile and more likely to reset silently after an update. A tool that worked for years can stop working after a point release, and the user has no signal until the moment autofill fails on a login they need.
At the same time the stakes compounded. A vault is now the root of a person's entire digital identity — dozens to hundreds of accounts, MFA seeds, recovery codes. When it becomes unreachable, the blast radius is everything, and the user discovers it at the worst possible time rather than in advance.
So the platform beneath these tools got shakier, the consequence of a silent failure got larger, and the vendors' own reliability has not kept pace — which is the precise gap where a user will pay for a second layer of assurance.
The wedge
Do not build another password manager. Build the layer that tells people their manager is working — and gets them back in when it is not.
- A lightweight monitor that periodically verifies the extension is alive and autofill actually fires, and warns the user before they hit a broken login rather than during one.
- A vendor-independent, encrypted emergency copy of the essentials — credentials, MFA seeds, recovery codes — so a lockout is an inconvenience, not a catastrophe.
- Guided recovery playbooks for the common lockout paths across the major managers, turning "no recovery route" into a checklist a non-expert can follow.
Start as the tool a user installs the day autofill breaks — and becomes the quiet insurance cautious users keep running before it ever does.
Risks and honest caveats
- You are adjacent to the most sensitive data there is. Any emergency copy of vault essentials must be local-first and zero-knowledge, or the cure is worse than the disease. That bar is the price of entry, not a feature.
- The vendors could fix reliability and shrink your wedge. Reliability regressions are bugs, and bugs get fixed. The durable product is the always-on assurance and recovery layer, not a one-time patch for today's broken extension.
- Trust is the whole sale. "Let me sit next to your password manager" is a high-trust ask. It is earned with open formats, auditable local operation, and no cloud custody of secrets — not marketing.
- Willingness to pay spikes at the moment of failure and fades after. People buy when locked out and forget when safe; price and position for the always-on version, not the rescue.
How to validate this further
Read the Security threads in the Pain Point Browser and test the reliability-and-recovery framing with the Idea Validator. Related: a portability layer for credential vaults.