From your AI assistant
Add Linnia to Claude, ChatGPT, Cursor, VS Code or Codex as an MCP connector. One connection covers the LinkedIn accounts you choose. See MCP overview.
What is Linnia
Linnia lets an AI assistant such as Claude or ChatGPT work on a rep’s LinkedIn: research people and companies, send connection requests, reply to messages and follow up. You ask in plain words; Linnia does the sending.
What makes it safe is that the assistant never touches LinkedIn directly. Every action goes through Linnia, which paces it like a person would, inside the rep’s working hours and well below LinkedIn’s limits, and, if you want, waits for a person to approve it first.
Three ways to use it
Section titled “Three ways to use it”From the dashboard
At app.linnia.ai you connect accounts, approve messages, watch activity, set limits and manage your team.
From your own code
The REST API at https://api.linnia.ai and signed webhooks connect Linnia to your CRM and tools. See the API introduction.
What it can do
Section titled “What it can do”- Research: search people and companies (with Sales Navigator filters when the account has it), read profiles, company pages, posts and the inbox.
- Outreach: connection requests with a personal note, messages and InMail, reactions, comments, posts, endorsements, Sales Navigator leads, accepting and withdrawing invitations.
- Follow-up: automations such as “when they accept, wait two working days, then send a message”, which stop the moment the person replies.
- Team memory: a shared timeline per person across every rep’s account, plus notes, so two colleagues don’t contact the same person.
- Morning brief: replies that need attention in priority order (each reply is triaged as interested, meeting, referral, not now, not interested, objection or out of office), new connections ready for a first message, and what is waiting for approval.
What keeps accounts safe
Section titled “What keeps accounts safe”- One action at a time per account, with random human-like gaps, only inside working hours.
- Daily and weekly budgets below LinkedIn’s thresholds, and a 15-day warm-up for new accounts.
- An early-warning health ladder that slows or pauses an account when LinkedIn pushes back, before it turns into a restriction.
- Guardrails that block the things that get accounts flagged: links in connection notes, the same text sent to many people, low-quality cold invitations, people on the do-not-contact list, and people a colleague contacted in the last 30 days.
- Approvals: new accounts start in approve mode, so nothing is sent until a person says yes.
Read more in Safety and limits.