For Replit's Replit
Stuck in Replit? Press the dot.
Replit gets you to running fast. Relay gets you to production-ready when the prototype starts being used.
When Replit needs a person.
01
When the agent's stuck on the same error
Replit's agent is best-in-class for a starting point. When it's circling a single error for the fifth iteration, a Relay engineer joins and resets the loop.
02
When you need to deploy outside Replit
Replit deployments are great for prototypes; for production traffic with a custom domain, observability, and a real database, the engineer can move you to Vercel + Neon (or wherever) without losing the work.
03
When auth needs to be real
Replit's quick-start auth is a good starting point. When you need SSO, RBAC, or a session model that scales, the press surfaces a person who's done it.
Common questions
- Can the engineer work directly in my Repl?
- Yes via screen share, or you can grant temporary read-write access to the Repl. Either way, the change history stays under your account.
- What if I want to leave Replit?
- Common move at the production cutover. The engineer can mirror the Repl to a Git repo, set up CI on GitHub Actions, and deploy to Vercel or Render.
- Do you support Replit's database (Replit DB)?
- Yes for prototyping. For production, the engineer typically migrates to Postgres (Neon, Supabase, or self-hosted) — the press covers the migration.
Related reading
- Field notes · 9 min readAnatomy of the handoff: what we learned shipping the first 1,000 sessionsA median match in seconds isn’t a marketing number; it’s a system constraint we engineered toward and missed by 16 seconds in the first month.
- Research · 17 min readWhen does an AI build want a person? A study of 4,200 stuck momentsWe logged every press in the first quarter of the private beta. Four patterns explain almost everything.
- Essay · 11 min readThe irreducibly human moment in softwareAn argument for why software engineering doesn’t shrink in the age of AI, it sharpens, and moves to a new place in the build.