One call in. Weeks of safe outreach out.
Agents speak in batches. LinkedIn speaks in limits. The queue engine holds the difference for as long as it takes.
Every step, in one builder
Invites, messages, InMail, profile visits, likes, delays and conditions, all as steps.
Build the sequence in the app, or let your agent build it over MCP. The queue engine works out the safe pace either way.
Tune the limits, per profile
A daily cap per action type, set per sender and raised as an account warms up. The queue never goes past it.
Work runs inside the hours you set, spread through the day rather than in bursts, each sender on its own proxy. An account that works when its owner works is an account that keeps working.
See which playbook is working
Accept and reply rates by day, per sender and per campaign. Two playbooks against the same audience stop being a matter of opinion: one of them has the better curve, and you can see which.
The same numbers read over the API, so the agent that built the sequence can judge it too.
Eduard
StephanA queue for every contact
Give an agent hands on LinkedIn and you inherit all of it: caps, cooldowns, events to watch, steps that must not run twice. That is the part we hold.