Meetlane

Inspiration

Lead databases are powerful, but they make it too easy to confuse more filters and more automation with better outreach. Meetlane starts from a simpler promise: describe who you want to reach, review one blocker at a time, and keep a human in control before anything reaches Gmail.

What it does

Meetlane moves a lead through Fit → Company → Evidence → Contact → Review → Gmail. Each stage exposes one unresolved decision and one safe next move. Sources and paid credit costs stay explicit. Contact unlock is intentional, drafts remain editable, and the final handoff opens Gmail for the user—Meetlane never sends automatically.

How we built it

During Build Week, Codex—powered by GPT-5.6—helped inspect the real Next.js product, isolate a release branch, restore a public sanitized judge path, tighten auth and analytics boundaries, add regression tests, and validate the complete workflow in Chrome. Supabase provides the production data layer, while the public sample uses local state and no external providers.

Challenges

The largest release risk was operational, not visual: main and staging had a substantial migration delta. We chose not to force a last-minute production deployment. The working production site remains stable, the qualifying code is in a reviewed draft PR, and judges have private repository access.

Accomplishments

• A coherent one-blocker workflow from Company through Gmail
• Human review as a real product boundary
• A no-login, contact-safe interactive sample
• 1,428 automated tests plus typecheck, build, and browser validation
• Honest documentation of current evidence-yield limits

What we learned

A trustworthy AI workflow needs visible limits. Codex and GPT-5.6 were most valuable not just for implementation speed, but for catching cross-layer release risk and keeping the submission grounded in what the product actually does.

What's next

Review the production migration set, promote the public sample safely, and improve evidence yield before expanding paid research.

Built With

• openai-codex