You wrote a clean Boolean string, but you know it is leaking candidates. The title you typed is one of five a person could carry, the skill has a French spelling and an acronym, and half your target market lists everything in another language. Widening that by hand — synonym by synonym, group by group — is the tedious part of every search. This tool does that pass with you.
Say Hello (or Bonjour to work in French) to lock your session language, then either paste a Boolean string or just describe who you’re looking for in plain words — find me salespeople in Quebec works as well as a full query. The chat splits the input into groups — titles, skills, locations, employers — and walks one group at a time. From a plain-language brief it infers the groups for you (titles, deduced skills, nearby cities, seniority) and builds them out with you. Give it "data scientist" and it comes back with the obvious neighbours plus inclusive forms, acronyms and translations: ("data scientist" OR "data engineer" OR "machine learning engineer" OR "ML engineer" OR "scientifique des données"). You validate each step before it moves on, and at the end it rebuilds the whole thing in a copy-ready block you paste straight into LinkedIn, Recruiter or GitHub.
The honest part: it never locks a group without your say-so. It suggests, you decide — so if it over-expands a tight title or guesses a synonym wrong, push back and tell it to drop the term or skip the group. Treat the output as a starting draft you trim, not a finished string. It runs on Workers AI (Qwen3 30B), no signup, and your queries aren’t logged.