DisperseAn AI that knows who's worth messaging on LinkedIn: and writes like you'd actually write it.
https://disperse.turtlelabs.io/
Product walkthrough of the Disperse dashboard: campaign setup, lead qualification, and the message review queue.
0:00 / 0:08
The sequencer you've seen 100 times
Subject: Quick question re: ScaleVector
Hi {{First_Name}},
I came across your profile and was impressed by what you are building at {{Company}}. Do you have 15 minutes this Thursday for a quick call?
Template detectedNo verifiable signal
Disperse draftSignals verified
Role change: 3 weeks agoShared employer
Hi Jordan, saw you moved into the Head of Ops role a few weeks back. Curious how the transition's going, especially since we overlapped at the same company a few years ago.
Every claim checkedHeld for your review
Failure modes of automation
Why generic LinkedIn automation burns your account and your reply rate.
Keyword-filter spam, sent to everyone identically
Sequencers dump hundreds of matches from a saved search into one sequence bucket. A staff engineer and an entry-level SDR get the same message with a name swapped in, and recipients spot the merge tag instantly.
Outcome: Ignored, or reported as spam
Cloud bot farms trigger account checkpoints
Tools that route your outreach through a separate machine, on a datacenter IP that looks nothing like you, are the exact pattern LinkedIn's fraud heuristics are built to catch.
Outcome: Checkpoints, restrictions, or a ban
Real personalization doesn't scale by hand
A genuinely researched message gets real replies. But reading someone's recent posts, job history, and shared connections takes real time per lead, and nobody has hours a day for it.
Outcome: Pipeline starves, or you burn out
How it's built
Built to behave like one careful person, not to disguise a machine as one.
Disperse is not a bulk sender with a language model bolted on. It is a screening and qualification engine that happens to draft the message at the end.
Active client node
Local browser execution
What it does: Runs every action inside your real, already-signed-in Chrome session, using your own cookies and your own IP.
Why it matters: There is no separate machine impersonating you, so there is nothing about your account's behavior for LinkedIn to notice as foreign.
Your own session, your own address
Pre-send filter
Qualification and pruning
What it does: Reads a lead's full profile, recent activity, and company signals before anything is queued to send.
Why it matters: Most imported leads were never going to reply. They get dropped here instead of spending a connection request.
Fewer requests, better targets
Anti-hallucination
Signal-grounded personalization
What it does: Anchors every specific claim to something observable: a role change, a shared employer, a mutual connection.
Why it matters: If there is no verifiable hook, the draft is flagged rather than padded with invented flattery.
Reads like you noticed, because it did
Zero blind sends
Human-in-the-loop controls
What it does: Routes drafts into a review queue where you can approve, edit, or reject before anything leaves your account.
Why it matters: You never find out after the fact what your AI said to a prospect. You set how much runs unattended.
Nothing sends that you would not have sent
Adaptive throttle
Defensive pacing
What it does: Applies randomized delays, a warm-up ramp on new accounts, and a hard cap on unanswered follow-ups.
Why it matters: It reads and clicks like someone with other things to do today, not a script running at machine speed.
Paced like a person, not a scriptBacks off on any rate-limit signal
Why Disperse
Three assumptions this category runs on, and what we do instead.
The assumption
Moving automation onto a separate cloud machine, away from your own browser, makes it safer.
What Disperse does instead
Runs through your own browser, on your own LinkedIn session, the same way you'd use LinkedIn yourself.
Why that’s the safer bet
A second machine acting as you is still a second machine acting as you, no matter how well it's disguised. There's nothing about your account's behavior for LinkedIn to notice as different, because nothing about it is different.
The assumption
More automation, more volume, more messages sent, is how you get more replies.
What Disperse does instead
Qualifies who's actually worth contacting before it sends anything, and learns from what actually gets replies.
Why that’s the safer bet
Fewer, better-targeted connection requests protect your account standing and your reply rate at the same time. Volume for its own sake is the thing that gets accounts flagged.
The assumption
Full autonomy, an AI that runs entire conversations with no human checkpoint, is the end goal.
What Disperse does instead
Flags anything it's uncertain about for you to review before it sends, and you decide how hands-off you want to be.
Why that’s the safer bet
You never end up finding out what your AI said to a prospect after the fact. You set the level of control, from fully reviewed to fully automatic.
How it works
From cold setup to qualified conversations in four steps.
- Step 1
Pair your LinkedIn
Install the extension and sign in once. From then on, Disperse acts through your real session: your cookies, your IP, your pace.
One-time pairing - Step 2
Set your voice, once
One core claim, stated exactly how you'd say it. The agent is only allowed to use your words, not improvise new ones.
Your words, not the model's - Step 3
Review, or trust the tier
High-value leads route to your inbox for a look before anything sends. Lower tiers can auto-send once you trust the pattern.
Per-tier review controls - Step 4
Track every reply
Accepts, replies, and state changes land in one thread, matched to the same pipeline you can watch live.
Replies land in your inbox
In production
Live production figures · updated hourlyThis has actually run, at real scale, already.
Profiles evaluated
0+
Each one checked against real qualification criteria before it could count as a lead.
Qualified leads surfaced
0
Passed multi-factor qualification and signal validation.
Qualification rate
22.0%
The rest were filtered out rather than contacted.
Most outreach tools report how many people you reached. The number that matters here is the one in the middle: out of everything Disperse looked at, only 22.0% was ever worth contacting, and the rest were dropped before they cost you a connection request.
The human checkpoint
This is what lands in your review queue.
Names and companies below are illustrative. The checks running underneath are exactly what every real draft goes through before it reaches you.
Follow-up · Alex RiveraAwaiting your review
Hey Alex, thanks for the reply! Since you mentioned the backlog your ops team's been fighting, we built the routing layer three teams like yours are already running on. Worth a 15-minute look, or should I check back in a few weeks?
Core claim intact
Draft generationClaim drift detected
“...we built the routing layer three teams like yours are already running at competitor scale”
Send blocked. Claim not in verified profile data. Regenerating.
Core claim, checked verbatim
The highlighted line is the one fact your campaign is built on. If a draft can't include it exactly, it gets flagged, not reworded.
No unfilled placeholders
A draft that still says {first_name} never reaches your queue. It's caught and regenerated before you see it.
Nothing sends on its own
Every opener and follow-up waits here until you edit it, approve it, or reject it.
Pricing
Founding workspace rates
Locked in for as long as your workspace stays continuously subscribed. Not a discount code that resets.
Individual
For one person running their own outreach
Regular ₹7,499/mo
₹2,999/ month
Billed monthly · Cancel anytime
- 1 connected LinkedIn account, on your own session
- Every claim of shared history verified before it's drafted
- Full qualification and pacing engine
- Human-in-the-loop review queue for every draft
- Pacing guardrails enforced in code, not switchable settings
Start your 14-day trial
A card is required. Nothing is charged during the trial. On day 15, ₹2,999/month bills automatically unless you've cancelled.
Team
RecommendedFor teams running several accounts
Regular ₹17,499/mo
₹6,999/ month
Billed monthly · Founding rate locked
- 3 connected LinkedIn accounts, each on its own session
- Everything in Individual, for every seat
- Shared target lists across your team
- Team-wide pacing governance, so no account exceeds a safe rate
- Priority routing through the review queue
Start your 14-day trial
A card is required. Nothing is charged during the trial. On day 15, ₹6,999/month bills automatically unless you've cancelled.
Questions
Direct answers to the questions that actually matter.
How is this actually different from the sequencer tools I've already tried?
Most tools in this category either run a script through a browser extension, which behaves nothing like a real person's pace, or move everything onto a separate cloud machine and treat that distance as the safety feature. Neither actually solves the problem. Disperse runs through your own browser and your own LinkedIn session, decides who's worth contacting before it sends anything, and flags anything uncertain for you to check. It's built to behave like one careful person, not to disguise a machine as one.
Will this get my LinkedIn account restricted or banned?
Does it actually sound like me, or like an obvious AI template?
Do I have to approve every single message before it sends?
What happens to my LinkedIn login details?
What if I'm not sure how to target the right people?
Stop sending the same automated message everyone already ignores.
Run your outreach through an agent that qualifies before it sends and writes from something real about the person, not a merge field.
A card is required. Nothing is charged for 14 days. Runs on your own browser.