Guides
How the work gets done.
Each guide follows one specialist through one job, from a prospect to a qualified reply. Read them to understand what the team does before you approve its first send, or to run the same play by hand.
Outbound playbooks
Signals
Find LinkedIn buyer-intent signals
What a real buying signal looks like on LinkedIn, and how the Signals specialist turns one into a reason to reach out.
Read the guideICP Scoring
Define and train your ICP
Turn a vague ideal customer into a scored definition, then train ICP Scoring on your own reviews.
Read the guideVoice Writer
Write outreach from a LinkedIn signal
From signal to first message in your own voice, without templates.
Read the guideDirector
Trace leads through outbound
Follow a prospect from list to qualified conversation and know which step earned the reply.
Read the guideReply Qualifier
Handle LinkedIn outreach objections
The replies that stall a conversation, and how Reply Qualifier sorts the ones worth your time.
Read the guideSafety Center
LinkedIn connection limits, explained
What LinkedIn publishes about invitation limits, what it does not, and how to stay under them.
Read the guideRun it from your own agent
The same six specialists over the Zevari LinkedIn MCP and REST API.
Connect Claude Code or Codex in 60 seconds
The LinkedIn MCP setup page: one command, OAuth, first three prompts.
Connect OpenClaw to LinkedIn
Workspace token, MCP or REST wiring, and a first safe find-score-stage-approve loop.
Connect Hermes Agent to LinkedIn
Register Zevari in the Hermes config and run the same approval-gated loop.
LinkedIn MCP prompts
Copy-paste prompts that drive real outbound sessions from Claude or ChatGPT.
Or let the team run the play.
Tell Zevari who you sell to. You train the six specialists on your ideal customer and your voice, then switch them to Autopilot.