We built Candidfy because some conversations never happen.
The hygiene concern nobody will raise. The feelings nobody will admit. The worry about a friend's drinking that everyone sees and nobody says. The feedback that would genuinely help someone — that stays unsaid because the social cost of saying it directly is too high.
These are not small conversations. They are often the most important ones a person could have. And they do not happen — not because people do not care, but because the mechanism for having them safely does not exist.
Candidfy is that mechanism.
What we built
Candidfy lets you send an honest message to anyone — anonymously. You write what you actually mean, without filtering. The AI rewrites it for clarity and respect, preserving every honest observation while removing cruelty. The recipient receives a private one-time PIN. They read the message, can reflect, and can reply.
A real conversation — one that would never have started otherwise — becomes possible.
The rewrite principle
Preserve the truth fully. Remove the cruelty — not the honesty.
This is the design principle that governs every rewrite. Candidfy's AI does not water down hard truths. It does not replace "you have a serious drinking problem" with "I have noticed you enjoy socialising." It makes the honest thing land — in a form the recipient can actually receive.
The difference between honesty that helps and honesty that harms is almost never the content. It is the framing.
The privacy covenant
Your identity is known only to Candidfy. It is never shared with message recipients, third parties, advertisers, or data brokers. This is not a policy that could change at the next board meeting. It is an architectural commitment: identity and messages live in separate encrypted databases that are never joined in production.
No ads. No data sales. Revenue comes only from users who find enough value to pay for it.
What we are not
Candidfy is not an anonymous harassment tool. Every message passes through a four-stage safety pipeline before delivery. Threats, stalking language, coercive control, and obsession patterns are blocked. Senders who attempt to circumvent the system are flagged and suspended.
Anonymity protects the speaker — not the harmful. The distinction matters and it is built into the architecture.
The use cases we care about
Contact
For general enquiries: hello@candid.app
For support: support@candid.app
For press: press@candid.app
For security concerns: security@candid.app