IronClaw: Local AI CRM for Lead Enrichment and Outreach

Run a full AI-powered sales pipeline on your Mac -- scrape founder directories, enrich contacts with LinkedIn data, send personalized outreach, and track conversions -- all locally with IronClaw.

What it does

IronClaw is an OpenClaw fork purpose-built for CRM and outreach automation. Instead of building a CRM layer on top of OpenClaw manually, IronClaw ships with it pre-built:

• Scrapes founder directories like YC batches, pulling names, companies, LinkedIn profiles, and education into a local DuckDB database
• Enriches contacts with LinkedIn profile data, email addresses, and company details -- using your existing Chrome profile for authentication
• Generates personalized outreach messages tailored to each contact's background and company
• Runs follow-up sequences on cron schedules (LinkedIn connection request, then email, then second email at 3-day intervals)
• Tracks pipeline stages from New Lead through Contacted, Qualified, and Converted with live analytics

How it differs from OpenClaw

OpenClaw is a general-purpose AI agent gateway -- you can build anything, but you design the structure yourself. IronClaw narrows the scope to one workflow done well:

	OpenClaw	IronClaw
Scope	Multi-agent orchestration, 15+ messaging channels	CRM, lead enrichment, outreach
Setup	Flexible, you build your own pipeline	Pre-built dashboards and pipeline stages
Browser	Playwright skill for scraping	Uses your real Chrome profile (cookies, sessions)
Database	Memory files	DuckDB with structured tables (contacts, companies, founders)
Platform	Linux, macOS, Windows (WSL)	Mac-first

Setup overview

1. Install IronClaw: curl -fsSL https://ironclaw.sh/install.sh | bash -- opens at localhost:3100
2. Connect your Chrome profile -- IronClaw detects active auth sessions (LinkedIn, Gmail, GitHub)
3. Import existing contacts from Google Drive, Notion, Salesforce, HubSpot, or CSV
4. Define your outreach sequence in the SOUL.md: target criteria, message templates, follow-up cadence
5. Set up cron schedules for enrichment runs and follow-up sequences

LLM and tools

Uses Claude (or any OpenAI-compatible endpoint via OpenRouter, Ollama, or vLLM) for generating personalized messages, analyzing lead quality, and summarizing pipeline status. DuckDB handles structured data queries -- ask questions in plain English and IronClaw translates to SQL. Chrome profile integration provides seamless LinkedIn and email access without re-authentication.

Tips

• Start with a small batch (20-30 contacts) to tune your outreach message templates before scaling
• Review generated messages before enabling auto-send -- the AI adapts to your style over time but needs initial feedback
• Use the analytics dashboard to track reply rates per message variant and adjust your approach
• Set respectful sending limits -- LinkedIn has daily connection request caps, and aggressive outreach gets flagged
• Export pipeline data periodically; DuckDB files are portable and easy to back up

Source

Based on IronClaw -- AI CRM, hosted locally on your Mac (2026), OpenClaw vs Ironclaw Comparison (Feb 23, 2026), and DenchHQ/ironclaw on GitHub (2026)