Four agents, one workspace, your rules
Sourcing, screening, ATS and scheduling agents share the same context, the same guardrails and the same audit trail. You approve what leaves the building.
Each agent has one job and one set of limits
Sourcing agent
Maps the market for a role — target companies, adjacent titles, recent movers — then checks your ATS and do-not-contact list before anyone is touched.
Screening agent
Runs the structured screen in your voice — notice period, comp, must-haves — and feeds every answer back into the fit score.
ATS agent
Reads and enriches your records, surfaces past applicants who fit the role you opened today, and writes notes back where your team already works.
Scheduling agent
Books qualified, interested people into your real availability, sends prep notes, and hands the thread to a human without losing context.
From one brief to interviews on your calendar
Launch your own agents for the recurring work
Write the instruction, set the cadence, review the output. They inherit the same tools, guardrails and audit log as the four core agents.
Reporting
Pipeline movement, time-to-first-reply, cost per hire — delivered where your team already reads.
Talent market watch
Who is hiring, who is cutting, where comp moved in the roles you own. Weekly, with sources.
Nudges
Chase quiet candidates twice then stop. Remind interviewers about missing feedback.
Rediscovery
Re-score the ATS on every new req and resurface people you already met.
Scorecard audit
Flags criteria nobody uses and bars the market cannot meet at your budget.
Anything recurring
If you can describe the job in a sentence, it can run on a schedule and report back.
A score you can argue with
Every reply, every rejection you log and every note from a screening call moves the score — and the reasoning is written out, line by line, so your team can disagree with it.
Agents that ask permission, not forgiveness
Gates per hunt, attribution on every action, and an export your legal team can read.
Approval gates
First send, follow-ups, scheduling, ATS writes — turn each gate on or off per hunt.
Read-only audit log
Every action attributed to an agent or a person, with the reasoning attached. Exportable.
Humans decide
Agents qualify and organize. Interviews, judgement and offers stay with your team.
Or run the whole thing from Claude
One connector, two minutes, no code. Source, message and check pipeline from any AI that speaks MCP — or install the official app in ChatGPT.
See the MCP page
Claude
ChatGPT
Le ChatHow the agents work
Want to see it on one of your open roles? Book a call and ask us directly.