MimikFlow

AILinkedIn prospecting and setting that gets real replies

Visit Website
April 16, 2026 Launching MimikFlow: $500 MRR in 2 weeks with the AI LinkedIn tool I built because I hated every other one

I've spent 10+ years building AI systems as a data scientist, and I got tired of watching the prospecting industry use generative AI to do the exact opposite of what it's good at.

We finally have models that can read context, adapt tone, and write like humans — and the entire LinkedIn outreach industry is using them to spray 1,000 "Hi {firstName}, I came across your profile" templates per day, faster than ever.

So I built MimikFlow. Here's where I'm at.

What it does

MimikFlow is an AI LinkedIn prospecting tool that actually reads each prospect's profile, posts, and activity, then writes messages that sound like a human wrote them — not a script. Connection requests, multi-step follow-ups, inbox management, all in one pipeline view.

No Chrome extension. No scrapers. Runs on the official LinkedIn API via Unipile, so nobody wakes up to a banned 5-year-old account.

The numbers (2 weeks in)

  • $500 MRR

  • 5 paying users at €197/month

  • 12 in active trial

  • 7-day free trial, no credit card upfront

  • On the users' campaigns so far: 54% reply rate, 15% booked on Calendly

That reply rate is the number I care about most. The industry average sits around 3–10%. When a tool writes messages that don't feel AI-generated, people actually reply.

Build timeline

1 month of focused dev to get MimikFlow where it is today — but that's misleading. I spent the 3 months before that prototyping the whole thing in n8n: the AI message logic, the sequencing, the LinkedIn API orchestration. Once the logic was validated with real outreach data, rewriting it as a proper SaaS was mostly plumbing.

If you're hesitating between "build the MVP in n8n" vs "go straight to code" — build it in n8n. I would have wasted 2 months if I'd started in Next.js.

Stack

Next.js, Supabase, Stripe, Unipile for LinkedIn API, Coolify on Hostinger for deploy. Claude Code handles most of the feature work now.

What I got wrong

Two things, honestly:

  1. I priced too low at first (€97). Moved to €197 and conversion didn't drop — the people who care about reply rates aren't price-sensitive.

  2. I underestimated how much the "no account ban" angle would matter. Half my trial users came in specifically because they got burned by a Chrome extension tool before.

What's next

Multi-inbox management, team plans, and an Instagram DM version (the same AI logic, different channel). Also shipping an affiliate program in the next 2 weeks.

Try it

7-day free trial at mimikflow.com. No credit card. I'm the one answering support, so if you sign up and something breaks, you'll hear from me directly.


Question for IH: for those of you who've launched a B2B SaaS in a crowded space — did you lean into a sharp contrarian angle ("less volume, more replies" in my case), or did you try to compete on the same axis as incumbents? Curious how that played out for you.

1 Comment

  1. 1

    Hey, your app caught my attention — nice work.

    I think the product can reach a much larger audience with proper distribution strategy.

    We focus on helping apps scale installs and reach relevant users.

    Happy to connect if this is relevant.

About

10 years in AI/ML, and I got tired of watching generative AI turn LinkedIn into a spam factory of "Hi {firstName}" templates. MimikFlow is the opposite bet: fewer messages, full setting, meetings booked easily by AI