How it works

Seven steps, in order. The second one always runs before the third — that sequencing is the point.

1. 01

   You describe it in your own words

   A chat, not a form. No dropdowns to squeeze your situation into, and no character limit on the part that matters.

2. 02

   A safety screen runs first — every message, every time

   Before anything is categorised, every message is screened for clinical risk. If something reads as crisis, we stop there and show real resources for your region. Nothing about that counts against your account.

3. 03

   We show you what we understood

   Topic, stage, intent — extracted and then handed back to you to correct before it's saved. If we got it wrong, your correction is the version that counts.

4. 04

   A shortlist, with the reasoning shown

   Rule-based matching on topic, stage, shared language and region. Each candidate shows what it matched on plus one line in their own words. No photos, no bios, no scrolling.

5. 05

   You reach out, or you don't

   Pick one person or several. Dismiss anyone as 'not a fit' — silently, with no consequence for either of you. Unanswered requests expire after 7 days so nothing hangs.

6. 06

   A private conversation, at your pace

   In-app messaging only. Either of you can end it with no fault attached, hide it from your own view, block, or report — four separate actions, because a mismatch isn't an accusation.

7. 07

   Afterwards

   Both sides are asked whether it helped. That aggregates into a trust score. Repeated negative feedback goes to a human for review, never to an automatic ban.