Scrylo

The problem I kept running into

Every lead gen tool I tried had the same business model: pay monthly, your data lives on their servers, hit a limit and pay more. Apollo, Hunter, ZoomInfo; great products, genuinely useful, but the pricing assumes you're a company with a budget. If you're an indie founder, freelancer, or small agency doing outbound, you're paying $50–200/month for something you use sporadically.

I wanted to find B2B leads, score them, write personalised cold emails, and send them. I didn't need a team dashboard or CRM integration. I needed a tool that worked, kept my data local, and didn't charge me every month for the privilege.

So I built Scrylo.

What it does

Scrylo is a local lead intelligence tool. You run it on your own machine — Windows, macOS, or Linux. Your data never leaves your computer. There are no servers on my end because there aren't any.

The full pipeline in one flow:

1. Find leads*l — type a query like `"SaaS CEO London"` or `"fintech founders in Berlin"`. Scrylo searches via Serper (Google) and scouts GitHub profiles to surface real decision-makers with verified contact details. MX record validation runs in real-time so you only save leads with working email domains.

2. Score against your ICP— define your ideal customer profile in plain English. Scrylo scores every lead 0–100 based on role relevance and keyword match. Hot leads float to the top.

3. Generate personalised pitches — uses Groq (llama-3.3-70b) to write a unique cold email per lead based on their role, company, and your service description. No Groq key? Falls back to smart built-in templates automatically. No dead ends.

4. Send from your own SMTP— works with Gmail, Outlook, Zoho, Brevo, anything standard. Randomised delays, List-Unsubscribe headers, bounce tracking, daily send cap. Your sender reputation stays intact.

You also get a business card importer (drop a folder of event photos, it reads every card and imports them as leads), a full web UI at localhost:8080, and a CLI for power users who prefer the terminal.

---

What I'm not doing

- No LinkedIn scraping (violates ToS, gets accounts banned — not worth it)

- No storing your data anywhere — literally no servers

- No feature gating — one price, everything included

- No monthly fees, no per-lead charges, no API credits to buy

---

The pricing

$99 once. Lifetime updates included.

One licence key works for two users. If it doesn't run on your machine and I can't fix it, full refund within 7 days.

The only ongoing costs are on your side — a Serper API key for web discovery (free tier is generous, more than enough to get started) and optionally a Groq API key for AI pitches (also free tier, and the fallback templates work fine without it).

What I learned building this

The hardest part wasn't the lead finding — it was the email quality. Getting an LLM to faithfully follow a user's service description rather than improvising its own CTA took more iteration than I expected. The model would read detailed instructions and still default to "let's hop on a 15-minute call" regardless of what you wrote. Solved it by rethinking the prompt architecture entirely.

The second hardest part was name filtering. When you scrape company team pages, you get a surprising amount of garbage — page section headers, product feature names, navigation labels — all formatted like "Title Case Name". Built a multi-layer filter combining structural rules, suffix pattern matching, and an optional LLM filter for users who configure a Groq key.

Where it stands

- Working on Windows, macOS, Linux

- Web UI + full CLI

- GitHub and Serper discovery

- Real-time MX email verification

- ICP scoring

- AI pitch generation with graceful fallbacks

- SMTP sending with anti-spam built in

- Business card importer

What I'm looking for

Honest feedback from people who do outbound. Does the pricing make sense? Is there something obvious I'm missing? Would love to hear from founders, freelancers, or small agency owners who've dealt with the same frust

ration.

If you want to try it: scrylo.site

Happy to answer anything.

Here's a video showing the full pipeline:

https://youtu.be/Z-Y6Xa3BbB8?si=ItoIHB7VCLWmVN-i