I built an autonomous cold-outreach + CRM machine for ~$1/month (no SaaS)

Most cold-email stacks cost $100-300/month once you stack a sender, a verifier, a warmup tool and a CRM. I wanted to see how far you could get for the price of a coffee. Here's the build - and what actually moved the needle.

The constraint: ~$1/month, fully autonomous

The only paid piece is a single Zoho Mail Lite mailbox (~$1/month) on a separate subdomain, so the cold sending never touches the main domain's reputation or the transactional email. Everything else runs on a small VPS with scripts. No Instantly, no Smartlead, no Apollo.

The pipeline

1. Sourcing - pull real businesses that publicly list their email (open data), filtered to a tight ICP: micro/solo/new businesses across the US, UK, Canada and Australia.
2. Drafting - a model writes a short, personalized-by-trade email. Under 80 words, one soft CTA, compliant footer, real unsubscribe.
3. Warmup + sending - daily volume ramps 5 to 45 over ~2 weeks, with human-paced gaps. A file lock and a duplicate guard make double-sends impossible.
4. Monitoring - IMAP watches for replies (pause + alert) and bounces (auto-suppress).
5. Follow-ups - most answers come from touch 2-3, so non-repliers get 2 gentle bumps on a day-3 / day-4 cadence.
6. First-party data - every lead is tagged: replied, clicked, converted (matched against real orders), or ignored.

Three lessons that surprised me

• Deliverability beats everything. SPF + DKIM on the sending subdomain got a brand-new domain into the primary inbox on day one.
• Open tracking is a trap now. Apple Mail Privacy Protection auto-opens everything, so "opens" are noise. Track clicks, replies and actual sales instead.
• The bottleneck is rarely volume. It's whether the offer converts and whether you follow up.

Why I'm sharing this

I run Growtoria, an AI-native growth studio for B2B founders - we build these lean, automated growth systems (and sell plug-and-play versions as digital products). If you'd rather not wire it up yourself, we also publish a set of free tools you can use right now.

The takeaway isn't "cheap beats expensive." It's that a small, honest, well-warmed system that follows up and tracks the right signals will out-convert an expensive one that doesn't. Build the boring parts well.