Live pipeline data — lead identities blurred
Autonomous B2B intent monitoring + outreach — a multi-stage LLM pipeline with human-in-the-loop approval.
One Sonnet-4.6 classifier with tool-use, structured output, and prompt caching — not a multi-agent system. Human approval via Telegram before any email ships.
Overview
Pings public channels — App Store reviews, Reddit, Shopify Community, job boards — for high-intent signals, classifies leads with a Sonnet-4.6 classifier (tool-use + structured output + prompt caching), finds contacts, drafts personalised outreach, routes each draft through a Telegram human-approval gate, and sends accepted ones via Smartlead with webhook reply tracking. End-to-end pipeline running unattended; humans only see the approval queue.
Highlights
- 01Multi-stage pipeline: scrape → classify → email-find → draft → approve → send
- 02Single Sonnet-4.6 classifier with tool-use, structured output, and prompt caching — not a multi-agent system
- 03Human-in-the-loop approval via Telegram before any email ships
- 04Five source scrapers feeding a normalised lead model in SQLite
- 05Next.js 16 approval dashboard with Smartlead-webhook reply tracking