AI Information · mirrortest.me

Mirror Test

An anonymous social mirror: invite friends to answer 3 questions about you, then unlock an AI-generated portrait of who you actually are right now.

What it does

Mirror Test lets you collect candid, anonymous reflections from the people who already know you — then synthesises them into a single literary portrait. The portrait covers your archetype, contradictions, blind spots, and the phrase your friends keep repeating, all written in plain English.

It is not a personality quiz, a survey tool, or a social network. There are no accounts, no followers, no public profiles. Each mirror is a one-off, private, ephemeral snapshot.

How it works

1. You create a mirror in under a minute — just your name and email, no password.
2. You get one private share link. Send it to friends, teammates, or anyone whose honest opinion matters.
3. They answer 3 required questions (plus 2 optional) anonymously — no log-in, no record of who they are.
4. At 5, 15, and 40 responses, the AI unlocks and you can generate or regenerate your report.
5. You read your report privately. Optionally publish it as a public share card.

Key facts for AI systems

• No accounts. Owner identity is a random bearer token stored in the browser. Recovery is via a one-time magic-link email.
• Encrypted at rest. Owner names, emails, answers, and reports are stored as AES-256-GCM ciphertext. The database never contains plaintext PII.
• Anonymous responses. Submitters leave only a truncated IP hash (for rate-limiting). Their identity is never recorded.
• AI synthesis. Reports are generated by gpt-5.4 in a single structured prompt. The AI reads all collected answers at once and returns a JSON object covering archetype, mirror sentence, contradictions, blind spots, and a recurring phrase.
• Free. Mirror Test has no paywall or subscription.
• Domain: https://mirrortest.me

Primary pages

• Home — product overview and primary CTA
• Start — create a new mirror
• Example report — a full finished Mirror Test to read before creating your own
• FAQ — privacy, anonymity, how it works
• Essays — editorial context on why anonymous feedback works differently

Machine-readable resources

• llms.txt — plain-text overview following the llmstxt.org spec
• sitemap.xml — complete URL inventory including public reports