The Best LinkedIn Cold DM Software: A Practical Buyer's Guide

Cold DMs on LinkedIn are in a strange place. They still work — better than cold email in most B2B categories — and they are simultaneously the most abused channel in software sales. Everyone's inbox is full of "Hi {firstName}, I saw you're in the space". LinkedIn has responded by getting much more aggressive about automation detection, and the tools have responded by getting much more careful about how they send.

This guide covers what you actually need to choose between: the categories of tool, the specific products worth considering, the limits that keep your account alive, and the message structure that separates a 3% reply rate from a 25% one.

The three categories of LinkedIn outreach tool

Almost every product on the market is one of three things, and the differences matter far more than any individual feature

1. Cloud-based senders

These run on a remote server with a dedicated residential or mobile IP assigned to your account. You close your laptop and sequences keep sending. Because the IP stays consistent and the behaviour is throttled server-side, these are the lowest-risk option and the only category worth considering if you are running outreach at any real volume or across multiple accounts. Expect $79–$120 per seat per month

2. Browser extensions

These drive your actual browser session. Cheaper, often more flexible, and they only work while your machine is on and logged in. The risk profile is higher: LinkedIn can see the automation running in the DOM, and extensions historically get detected in waves. Fine for a solo founder sending twenty invites a day. Not fine for a team

3. Scrapers and enrichment tools

PhantomBuster, Clay, Apollo and similar products build and enrich the list. They are not cold DM tools, even when they technically can send. The correct architecture is almost always: scrape and enrich in one tool, send in another. Treating a scraper as your sequencer is the single most common reason accounts get restricted

The tools worth considering

Pricing below is indicative list pricing and moves constantly — check before you buy. What does not move is the shape of each product

How to actually pick one

Ignore feature grids. Three questions decide it

How many LinkedIn accounts are you sending from?

One account: almost anything works, and Waalaxy or LinkedHelper will save you money. Three to ten accounts: you need a cloud tool with a real unified inbox, and this is where HeyReach pulls ahead — it was built for exactly this and it shows in the pricing model. Ten-plus, agency-style: HeyReach or Expandi, and budget for someone to own it

Is LinkedIn your primary channel or a supporting one?

If email is the main motion and LinkedIn is a touch inside the sequence, buy an email-first tool with LinkedIn steps (lemlist, Smartlead-style stacks) rather than bolting email onto a LinkedIn tool. If LinkedIn is the main motion, do the reverse

Who is going to reply to the replies?

This is the question nobody asks and everybody regrets. A campaign that works generates conversations, and conversations need a human within a few hours. If you cannot staff that, do not turn on volume. A unified inbox with assignment and tagging is worth more than any sequencing feature

Safe sending limits in 2026

LinkedIn does not publish hard numbers and the thresholds are behavioural, not fixed. These are the conservative ranges experienced senders operate inside

• **Connection requests:** 15–20 per day on a new or low-activity account, up to 80–100 per week on a warm, established one. Weekly caps matter more than daily ones
• **Direct messages to existing connections:** 40–60 per day is comfortable; go higher only if reply rates stay healthy
• **InMails:** limited by your Sales Navigator credits — treat them as expensive and write them accordingly
• **Profile views:** keep automated views under a few hundred a day; spikes here are a common detection trigger
• **Warm-up:** ramp any new account over two to three weeks rather than starting at full volume
• **Acceptance rate:** if it drops below roughly 20%, stop and fix targeting — a low acceptance rate is itself a risk signal to LinkedIn

The other rule: never run two automation tools on the same account at the same time, and never run a cloud tool from one country while browsing from another. Inconsistent location is the fastest way to a restriction

Sales Navigator: do you need it?

For most serious outbound, yes — not for InMails but for the search. Boolean filters, saved-search alerts, headcount-growth and job-change filters, and the ability to pull a genuinely targeted 500-person list instead of a sloppy 5,000-person one. Cold DM performance is mostly a list problem, and Navigator is the cheapest fix for a list problem

The exception is if you already run Clay or Apollo with good data. Then Navigator is a nice-to-have rather than the backbone

The message structure that gets replies

Tooling gets you delivery. Structure gets you responses. The pattern that consistently outperforms is short, specific and asks for almost nothing

1. **Connection request: no note, or a five-word note.** Blank requests routinely outperform pitchy ones on acceptance. If you write a note, make it about them and keep it under 200 characters
2. **First message: one observed detail, then one sentence of relevance.** The detail must be something only true of them — a hire, a launch, a post they wrote. Not their industry
3. **No link, no calendar, no deck in message one.** Every asset you attach lowers reply rate. You are buying a reply, not a meeting
4. **Ask a question they can answer in ten words.** "Are you handling this in-house right now?" beats "Would you be open to a 15-minute call Tuesday?"
5. **Follow up twice, four to seven days apart, then stop.** Each follow-up should add information, not repeat the ask. Third follow-ups burn goodwill for a rounding error in replies

One more thing worth saying plainly: personalisation tokens are not personalisation. {firstName} and {companyName} were a differentiator in 2019 and are a tell in 2026. If your first line could be sent to a thousand people, it will perform like it was

What makes cold DMs work better: warm profile, warm audience

The uncomfortable finding that shows up in public teardowns and outbound case studies: the same message sent by someone who posts regularly outperforms the same message from a dormant profile, often by two to three times. The prospect clicks your profile before they reply. What they find there decides whether the reply happens.

Which is why outbound and content are not separate budgets. A founder posting twice a week is running a permanent warm-up campaign for every cold DM they send — and the people who already saw a post reply at a completely different rate How founder-led content compounds

If you want the audience side handled while you run outbound, that is the thing we do See how it works

Common mistakes that kill campaigns

• **Buying volume before validating the message.** Test on 50 people manually before automating 5,000
• **One sequence for every segment.** Different roles need different first lines; the tool makes this cheap, so use it
• **Sending from a half-finished profile.** Banner, headline and a recent post do more for reply rate than any sequencing feature
• **Ignoring the acceptance rate.** It is the earliest signal that your targeting is wrong
• **No CRM sync.** Conversations that live only in LinkedIn's inbox get lost the moment someone goes on holiday
• **Running the tool from a VPS in a different country to your usual login.** The single most reliable way to get restricted

Is LinkedIn automation against the terms of service?

Yes — LinkedIn's user agreement prohibits automated access and third-party software, and no vendor can change that. In practice enforcement is behavioural: accounts that look like humans and generate positive engagement are left alone, and accounts that spray are warned, then restricted, then removed. Treat every tool as something you use carefully and reversibly, keep your volumes conservative, and never build a business where one account restriction ends the pipeline

The shorter version

Use a cloud sender — HeyReach if you run multiple accounts, Dripify if you want the simplest thing that works, Expandi if you need advanced logic. Build the list in Sales Navigator or Clay. Stay under 100 connection requests a week. Write first messages that could only be sent to one person. Post consistently so your profile does half the selling before the reply. Staff the inbox

Frequently asked questions

What is the best LinkedIn cold DM software?

For teams running multiple LinkedIn accounts, HeyReach is the strongest option because it was built around multi-account sending and a unified inbox. For a single user who wants the simplest setup, Dripify. For advanced conditional sequences and personalised images, Expandi. For solo founders on a tight budget, Waalaxy or LinkedHelper

How many LinkedIn connection requests can I send per day safely?

Stay around 15–20 per day on a new account and no more than 80–100 per week on an established one. Weekly limits matter more than daily ones, and a low acceptance rate raises your risk regardless of volume

Can LinkedIn detect automation tools?

Yes. LinkedIn detects browser-based automation more easily than cloud tools with dedicated IPs, but all automation carries risk. Consistent location, conservative volumes, human-like timing and a single tool per account are what reduce it

Do I need Sales Navigator for cold DMs?

Not strictly, but it substantially improves targeting through Boolean search and filters like job change and headcount growth. Since cold DM performance is mostly a list problem, it is usually the highest-return part of the stack

Are cold DMs on LinkedIn still effective?

Yes, and typically better than cold email in B2B — but only with tight targeting, genuinely specific first lines, and a profile that looks credible when the prospect clicks it. Generic templated volume performs worse every year

Is free LinkedIn automation software worth using?

Free tiers such as Waalaxy's are fine for testing whether outreach works for your offer. They are not built for volume, they usually lack a proper inbox, and the risk profile is higher because they run in your browser