Approval-first LinkedIn outbound

Outbound that never embarrasses you.

ContactBrew warms prospects with real engagement written in your voice, then asks only when they already know your name. Every word waits for your approval. Nothing goes out under your name that you did not see.

Full pipeline on the free planNo credits, everSeats are free

100%
of messages human-approved, by default
≤ 80
connection requests per week, a hard ceiling
You
are the send button, by default
0
credits. The listed price is the price
How it works

Warm first. Ask last.

Cold DMs on day one get ignored because the recipient has never seen your name. ContactBrew runs the sequence a thoughtful human would run, at a pace a thoughtful human would keep.

1

Source & score

Leads are sourced nightly against your ideal customer profile and scored with visible reasoning. High-confidence leads enroll; the rest wait for your tap. Nothing enrolls itself.

we call this “Beans”
2

Warm in your voice

The AI reads each prospect's fresh posts and drafts comments trained on things you actually wrote, with hard rules against flattery, emoji, and pitching. You approve every one.

your voice is “The Blend”
3

Connect, then ask

A blank connection request lands after days of visible engagement. They accept because they recognize you. The DM comes later, and only after they do.

paced by the governor
4

Reply means stop

The moment someone answers, their sequence stops permanently and the conversation lands in one unified inbox. ContactBrew never messages a person who already replied.

enforced, not a setting
Why warming works

By the time you ask, you're not a stranger.

A day-one pitch arrives with zero context: your name means nothing, so the ask carries all the weight. A warm-first sequence spends two weeks putting your face and your thinking in front of the prospect, so the eventual message opens a conversation that already exists.

  • Engagement is real and attributed. No fake views, no bots
  • Blank invites, sent only after visible engagement
  • The ask is the last step, never the first
Familiarity before the ask
Illustrative: how the two approaches differ in kind, not measured data.
Warm-first sequenceDay-one cold DM
Day 1Day 7Day 14Day 21Cold DM, day 1: ignoredlikecommentconnectMessage, day 16
Today's queue
12 comments · 5 DMs · 34 likes
your call
DM batch: 5 drafted, staggered on send
“Your point about paid work trials matched something we measured. Happy to share the numbers if useful. No pitch.”
Pour all 5EditCopy instead
The approval queue

You are the send button.

Every draft arrives with the post it answers, the person it's for, and their score against your ICP. Approve, edit, skip, or copy it to your clipboard and paste it yourself. That escape hatch is always there and always free.

  • Comments reviewed one by one. This is where your judgment earns its keep
  • DMs approved as a batch, then staggered over hours so they never fire as a burst
  • Every edit you make trains your voice profile
  • Want full auto for an action type? You type its name to turn it on. A checkbox is too easy for something with your name on it
Safety

Built like the account it protects is yours.

No tool can make LinkedIn automation risk-free. We say that out loud. What a tool can do is refuse to behave like a bot on your behalf.

The governor

Eleven checks before any action executes: daily and weekly caps, gaps between actions, per-contact cooldowns, your active hours in your timezone, blackout dates. It has no bypass. Not for us, not for any plan tier.

The circuit breaker

Two consecutive failures or a single platform challenge halts the account entirely and notifies you. No automatic retry, no backing off into the wall. A human turns the kettle back on.

Hard ceilings

Limits below LinkedIn's tolerance, enforced regardless of your settings. You can lower any cap. There is no code path in the product that raises one.

Quality gates

A safety gate blocks drafts touching politics, religion, or your competitors, and a semantic duplicate check stops the AI from repeating itself. In auto mode these gates cannot be turned off.

The ceilings, concretely
Enforced per account. Lowerable by you; raisable by no one.
Connection requests
80 / week
Comments
40 / day
Likes
100 / day
All actions combined
150 / day

And the part competitors put in small print: we don't build stealth modes or detection evasion, because nothing survives a platform that fingerprints cadence and device consistency. Discipline is the only strategy that ages well.

The honest column

What the pricing page usually hides.

The most common complaint in this category is that a plan's listed price is not its real price. Compare the fine print, not the headline.

 ContactBrewTypical credit-based tools
Billing dimensionPer LinkedIn accountPer account + per seat + credits
Team seatsFree and unlimitedPaid per user
AI draftingUnmeteredCredit packs that run out mid-month
Human approvalThe default, for everything with your name on itOptional, often buried
ProxiesIncluded at Agency and aboveBring your own, $5 to $15 per account monthly
On the risk of automationStated plainly, engineered around“100% safe”, allegedly
Pricing

Priced like we mean it.

Billing is per connected LinkedIn account. Everything else, from seats to drafting to enrichment, is simply included.

✓ No credits, ever. That's a promise printed on the pricing page

Kettle

Free
forever · 1 account
  • The full pipeline, not a demo
  • Sourcing, scoring, drafting
  • Manual send · 10 drafts/day
Start free

Growth

$49 /account/mo
3 to 9 accounts
  • Everything in Solo
  • Team roles
  • Unified inbox across accounts
Start free

Agency

$749 /mo flat
up to 25 accounts
  • White label
  • Proxies included
  • Client-scoped operators
Talk to us

Unlimited

$1,499 /mo flat
unlimited accounts
  • Everything in Agency
  • Priority support
Talk to us

Seats free on every plan · AI drafting unmetered on every paid plan · Enrichment included · Cancel anytime

FAQ

Asked before buying.

What is ContactBrew?

ContactBrew is a signal-based LinkedIn outbound platform. It sources leads against your ideal customer profile, warms them with real engagement written in your voice, and only asks for the meeting once they already know your name. By default, every comment, connection request, and message waits for your approval before anything moves.

Will ContactBrew get my LinkedIn account banned?

No tool can make LinkedIn automation risk-free, and any tool that claims otherwise is lying to you. What ContactBrew does is enforce the discipline that protects accounts: hard ceilings below LinkedIn's tolerance (80 connection requests per week, 40 comments and 100 likes per day), human-paced scheduling with jitter, active-hours windows in your timezone, and a circuit breaker that halts everything at the first sign of a challenge and never retries into the wall. You can lower any limit; nothing in the product can raise one.

Does the AI send messages without me seeing them?

Not unless you explicitly turn that on, action type by action type, by typing the action's name to confirm. Out of the box only likes and profile views run automatically, the two actions with no content and no inbox. Comments, connection requests, and DMs are drafted for you and wait in an approval queue. That default is the product.

How is ContactBrew priced?

Per connected LinkedIn account, never per seat. Kettle is free forever (one account, full pipeline, manual sending, 10 drafts a day). Solo is $59 per account per month, Growth is $49 per account per month from 3 accounts, Agency is $749 per month flat for up to 25 accounts with white label, and Unlimited is $1,499 per month flat. Seats are free and unlimited on every plan.

Are there credits or usage limits on AI drafting?

No credits, ever. Not for drafts, not for enrichment, not for lookups. AI drafting is unmetered on every paid plan. LinkedIn's own ceilings naturally cap volume, which is exactly why we can afford to not meter it. The price on the pricing page is the price you pay.

How does the AI write in my voice?

You give it 10 to 20 comments and DMs you actually wrote, plus style notes. Drafts are generated against that profile with hard house rules: no flattery openers, no restating the post, no emoji, no pitching. Every edit you make in the approval queue is stored and feeds the profile, so it gets more like you over time. A semantic duplicate check blocks the AI from repeating itself, which is the failure mode every other AI commenting tool eventually suffers.

Can my team or agency use one workspace?

Yes. Seats are free and unlimited, roles separate owners, admins, operators, and viewers, and an operator can be scoped to specific LinkedIn accounts. An agency VA sees exactly one client's queue and inbox, nothing else. All replies land in one unified inbox across every connected account.

What happens when a prospect replies?

The sequence halts for that person immediately and permanently. It's enforced in the state machine, not by a setting someone can forget. Their reply lands in your unified inbox, assigned to whoever owns the conversation. ContactBrew never messages someone who has already answered.

Your name is on every message.
Nothing sends that you didn't approve.

Start on Kettle, free forever. Run the full pipeline with manual sending and zero platform risk, and upgrade only when it earns it.

Start brewing, free