Type anything. Get a screened shortlist.
Type who you are looking for, or point Sonarnote at the briefing call you already recorded. Either way it works out the requirements, searches LinkedIn, and checks every candidate against each one.
Type a sentence
Use a recorded brief
Built to be checked, not believed.
Sourcing tools compete on how many profiles they hold. Sonarnote competes on whether you can defend the shortlist to the person who has to interview it.
Evidence, not a match score
Every other tool hands you a percentage. Sonarnote shows which requirement each candidate passed or failed, so you can check the machine instead of trusting it.
Rules decide, not the model
The model reads profiles. Fixed rules decide who qualifies, and a claimed skill with nothing in the profile behind it counts as unmet.
Change your mind for free
Loosen the seniority, drop a skill, widen the location. Everyone already screened is re-checked without spending another profile from your plan.
Your thumbs steer the next pass
Thumbs up and down tell the next pass in that search what you are after. More like this takes someone you liked and looks for the same shape.
Talk to the search
The search is a conversation, not a form. Ask it to widen the field or explain why someone was passed over.
Next to the interviews
Shortlists attach to the job. Interviews with those candidates land in the same workspace as your transcripts, notes, and scorecards.
Candidate sourcing, answered.
How does AI candidate sourcing work in Sonarnote?
You describe the role, either by typing it or by pointing Sonarnote at a recorded briefing call. It extracts the requirements into a card you review, searches public LinkedIn profiles once you approve them, and checks every profile it finds against each requirement separately. What comes back is a shortlist where each candidate shows which requirements they met and which they missed, with the line from their profile behind it.
How is this different from AI shortlisting tools that give a match score?
A match percentage tells you nothing you can act on. You cannot see which requirement drove it, you cannot tell a candidate who missed one nice-to-have from one who missed the only thing that mattered, and you cannot repeat the reasoning to a hiring manager who disagrees. Sonarnote shortlists per requirement instead, so what you get is a table of claims you can audit rather than a ranking you have to trust.
How does automated candidate screening decide who qualifies?
The model reads each profile and returns a verdict per requirement with a supporting quote, but it does not decide who qualifies. That is computed by fixed rules. Location and years of experience are absolute unless you mark them negotiable. A candidate has to match at least half the must-have skills, because briefs routinely split one competency into six narrow keywords. Across seniority, field, and language a candidate can fall short on exactly one and still qualify. And if the model claims a skill it cannot cite anything specific for, the check is forced to a miss.
Do I need to write a boolean search string?
No. There is no boolean syntax, no nested brackets, and no keyword juggling. You describe the role the way you would explain it to a colleague, review the requirements Sonarnote pulled out, and correct anything it got wrong before the search runs.
Do I have to record a call to use it?
No. Typing a description of the role is the normal way to start a search. Recording the briefing call is an option that saves you re-typing what the hiring manager already told you, and it means the requirements come from what was actually said rather than from what you remembered afterwards.
Where do the candidate profiles come from?
Public LinkedIn profile data, accessed through third-party data providers that search public web sources. Sonarnote does not need your LinkedIn account, your LinkedIn Recruiter seat, or any connection to your network.
Does it give me candidates' email addresses?
No. Sonarnote is a discovery and screening tool, not an outreach tool. Each shortlisted candidate links to their LinkedIn profile and you reach out however you normally would. There is no contact reveal, no email sequencing, and no automated messaging.
How long does a search take?
Roughly a minute for a new search, and most of that is the profile provider rather than the screening. You do not have to sit on the page while it runs. The shortlist appears when the run finishes.
What counts against my plan?
LinkedIn profiles searched, not searches run. Searches are unlimited on every plan, and you are billed for the profiles the data provider returns for them. Starter includes 500 profiles per month and Pro includes 1,000. Re-screening candidates you have already paid for, after editing the requirements, costs nothing extra.
What if the shortlist is wrong?
Tell it. Add or remove skill and industry chips on the requirements card, or say what is off in the chat, and the candidates already screened are re-checked against the corrected brief. In practice the first shortlist is a way of finding out that the brief was vaguer than everyone thought.
Can recruitment agencies use it for client briefs?
Yes, and the briefing call is where it fits best. Record the call with the client, let Sonarnote turn what they said into a requirement list, and go back to them with a shortlist where every candidate is mapped against the requirements they gave you. Disagreements about fit become disagreements about the brief, which is a much shorter conversation.
Is candidate data handled properly?
Sonarnote is EU-based and GDPR-ready. Profile data comes from third-party providers searching public web sources, is held against the search you ran it for, and is deleted with that search. The privacy policy, linked in the footer, covers the detail on data sources and retention.
Your next brief is a shortlist away.
Describe the role, review what Sonarnote pulled out of it, and get a shortlist you can defend. Searches are unlimited on every plan.