Lead-gen
automation that
keeps your
account alive.
Watch the 2-minute
dashboard walkthrough.
MeridianA live screen capture of the dashboard: import, build, approve, converse — and the auto-pause moment.
Everything you need
to run safe outreach.
Find & import leads
Search LinkedIn by keyword or describe your ICP and let web-grounded research surface real people. A confidence scorer rejects wrong-person matches, and every lead is deduped by publicId.
Paced campaigns
Warm-up → connect → DM → AI reply. Every step is gated, capped, and jittered — message only after they accept.
Unified inbox & CRM
Realtime DM threads with labels, pipeline stages, private notes, and due-date reminders — wired straight into campaigns.
Control & insight
An approval queue with live policy-aware countdowns, plus analytics on acceptance, replies, and pipeline movement.
Every step, paced
and approved.
Compose warm-ups, connection requests, DMs, and AI replies in a Kanban builder. Each step waits for your approval and fires under the rate limiter — never in a burst.
Build rapport before any ask
Rate-capped views, follows, and likes warm the relationship first — quiet social proof on real endpoints, never DM spam, never a burst.
A paced connection request
One connection request at a time, spaced with jitter and capped per day — invitations trickle out on human pacing, never as a batch.
A human-approved DM
The message step only unlocks once they accept your request — then the DM waits in the approval queue until you review and approve it.
Goal-driven, then hands off
Drafts replies toward a goal, recognizes when the goal is met, and escalates to you when it's stuck — every reply still gated through approval.
From a list to a booked meeting —
in four steps.
Import
Search LinkedIn by keyword or describe your ICP. Confidence-scored candidates land in your list.
Build
Compose a paced campaign in the Kanban builder: warm-up, connect, DM, AI reply.
Approve
Review every action in the queue. Edit a payload inline, approve, or skip — nothing fires unseen.
Converse
AI drafts a goal-driven reply, you approve it, and the conversation moves toward a booked meeting.
Runs inside LinkedIn,
talks to your server.
An in-page widget lets you push members into a list without leaving the feed. Outbound runs from an extension in your own authenticated session — never replayed from a server.
Capture from the feed
A shadow-DOM widget on feeds, profiles, and search pushes members into a list — deduped on the way in.
queue.enqueue({
type: 'CONNECTION_REQUEST',
gate: 'awaiting_approval'
})
Sub-second updates over a push stream — new messages, acceptances, and gate unlocks arrive in seconds, no polling.
Account safety isn't a setting —
it's the architecture.
Multi-layer, human-in-the-loop defense: routing, rate-capping, detection, and a fail-safe pause. Bursts are structurally impossible — even when you batch-approve.
Dispatch-gated rate limits
Approve 100 actions and they still won't burst — rolling day/week caps decide how many actually fire.
Fail-safe auto-pause
One detection block freezes the whole account atomically and notifies you. Never a retry, never a push-through.
The LLM never sends on its own
Every AI reply flows through approval, caps, jitter, and business-hours pacing before it touches LinkedIn.
Runs in your session.
Your server, your data.
Paced, approved,
never bursting.
Every action waits for your call — capped, jittered, and human-approved before it touches LinkedIn.
Run it your way.
Your infra, your keys
- Run on your own server
- Bring your own LLM keys
- Single account
- Full source access
- Community support
We host it for you
- Managed hosting
- Guided session setup
- Priority support
- Approval queue + analytics
- Realtime inbox & CRM
For teams at scale
- Multi-account
- SSO
- Org-scoped isolation
- Dedicated support
- Custom contracts
Outreach that respects
the account it runs on.
Self-hosted, human-approved, paced for longevity — not throughput. Connect your session and stay in control.