How do you move a Replit app to your own server?

Step by step

1. 1

   Push the Repl to GitHub

   Use the Git pane in the Replit workspace to connect a GitHub repository and push, or download the project as a ZIP and push it from your computer. Check that the run command from .replit (or package.json) is recorded, because the new host needs it.

   From a ZIP

   unzip repl.zip && cd <project>
   git init && git add . && git commit -m "from replit" && git push -u origin main

2. 2

   Choose the shape of the host

   One web process (Express, Flask, FastAPI, Next.js): Dashboard → Deployments → New with the repository and a pack; the framework is detected. Several processes, a bot that must run all day, cron jobs, or anything that writes to local disk: Dashboard → Instances → New for a full-root instance and deploy the repository onto it, or install things by hand from the web terminal.

3. 3

   Re-enter the Secrets

   Replit Secrets are stored outside the project files and are not in the export. Add each on the Environment tab (set_env over MCP), or as an .env on the instance. list_env_vars shows the names you have set without the values.

4. 4

   Move the database

   A Replit PostgreSQL database: pg_dump it from the Repl shell and restore it into a managed Postgres from Hostwares (create_database over MCP gives you one; db_connection_info returns the connection string). The Replit key-value store has no equivalent: export it to JSON from a short script and load it into a table.

   In the Repl shell

   pg_dump "$DATABASE_URL" --no-owner -Fc > app.dump
   # then, from your computer
   pg_restore --no-owner -d "<new DATABASE_URL>" app.dump

5. 5

   Point the domain and verify

   Domains tab on the site, or the instance's domain panel, then create the record it shows at your registrar. The certificate is issued once DNS resolves. Hit an API route, let the app run for a day, and read the logs (get_logs) for anything it wrote to a path that no longer exists.

Questions