Opens in the morning and shows an agent real people publicly asking about their niche right now, scored for buying intent, with the tools to work them into clients.

Independent agents lose hours hunting for people who are actually in the market. Pickup flips that around. It opens in the morning and shows real people publicly asking about your niche right now, scored for buying intent and ready for you to reach out. You make the connection; the app does the finding. The whole product is built around one job, getting an agent in front of people who are ready to buy, with the busywork handled so the agent can spend the day on conversations.

I built it end to end as the sole developer: the data model, the scoring feed, authentication, the integrations, and the interface. The design principle was focus. Rather than a general CRM with a hundred features, Pickup does one thing well and wires AI into the repetitive parts instead of bolting a chatbot on the side.

The intent feed. A feed of live prospects scored for buying intent, so an agent starts each day on the warmest leads rather than a cold list.

The tools to work them. Contact management, documents and PDF generation, and QR codes for sharing, so a lead moves from first contact toward client inside the same app.

AI on the busywork. An assistant built on the Anthropic SDK that drafts and processes the repetitive parts of outreach and follow-up, placed where it saves real time.

The plumbing. Stripe billing, S3 uploads with presigned URLs, and transactional email, so the product runs as a real business tool rather than a demo.

Next.js 16 and React 19 with Auth.js, Neon Postgres, the Anthropic SDK, Stripe, AWS S3, and pdfkit. The outcome is a tool built around a single clear job, finding people who are ready to buy, with AI woven into the workflow instead of sitting beside it.