Skip to main content
You do not install @replohq/sdk yourself. Every Replo site already has it, and Replo keeps the version current.

​
How it gets there

The SDK is a published npm package, and Replo adds it to your site’s package.json and installs it for you. When a new version ships, sites move to it automatically. You do not need to bump a version or run an install to pick up a fix. That means two things when you are editing site code:
  • Import it without adding a dependency. The package is already present.
  • Do not pin it. Replacing the managed version with a fixed one stops your site from receiving fixes, and Replo will move it back on the next update.

​
Peer dependencies

The SDK expects the site to provide these. A site scaffolded by Replo already satisfies them: React Query is what the data loaders read from, which is why ReploProvider has to wrap your app before any loader renders. See Setup.

​
Checking the version

BUILD_METADATA reports the package version the running site was built against:
That is the value to quote in a bug report.

​
Working on the site locally

Clone the repository, install, and run the dev server as you would any Next.js app:
See Connect with Git for the clone URL and authentication.

​
Next steps

Setup reference

The provider, config, and script tags a site mounts once.

Data loaders

Render store data on a page.