Leafer
A lead generation system run by AI agents that find buyers who are already asking for what you sell.
- Lead generation SaaS, AI agents
- 2026
- Own product. Agent design, engineering, infrastructure.
- Next.js, TypeScript, Postgres, AI agents, Background workers
The problem
Small B2B teams spend hours on cold lists full of people with no current need. The buyers who are actively asking for help, in public, are scattered across forums, social feeds and map listings, and nobody has time to watch all of them.
What we built
- Discovery agents that read public posts on Reddit, X, LinkedIn and Maps every day and score each one for buying intent.
- A lead generation pipeline that enriches and verifies every contact before it reaches a salesperson.
- A drafting agent that writes a reply quoting the exact post it came from, held in an approval queue so a person signs off on every message.
- Per-customer quotas, usage tracking and guardrails, so agents stay inside a budget.
- Multi-tenant workspaces, billing, an affiliate program and an admin console.
- Background workers, EU-hosted Postgres and a gated CI pipeline with end-to-end tests.
Where it stands
Live in public beta. Discovery, enrichment, scoring, drafting and the approval queue run in production today; outbound sending is deliberately switched off until it has been reviewed.