Vibe-code inside your boundary.

Your people already build apps with Claude. sprig plants what they build in your own Azure tenant. Entra-secured by default, governed by your Ops team, reading corporate data with each viewer's own identity.

What is sprig?Security & trust

The artifact is easy. The connection isn’t.

An app vibe-coded in Claude can’t reach your corporate databases, internal APIs, or other governed data. It lives outside your boundary. A demo, not a tool. sprig is the deploy target that brings it inside. One prompt moves the artifact into your tenant, where Entra decides what it can reach, for the owner and for everyone they share it with.

Evaluating sprig?

Start with What is sprig, then see how it works and the evaluation FAQ.

Security review?

Trust model, the data-access principle, and the exhaustive list of permissions the installer creates in your tenant.

Installing sprig?

Requirements → Install → onboard builders. One command runs preview, install, deploy and smoke test.

Reference

The sprig CLI, platform API, skills and configuration.

How a publish happens

A builder describes an app to Claude and says “share this solution”. Claude uses the organization-provisioned connector and release-matched skills. Claude Code uses the same remote MCP connection.
Claude calls the platform’s publish tool through the gateway, with the builder’s own Entra identity on every call.
The gateway stores a new immutable version of the app’s bundle and flips it live at <app>.apps.<yourdomain> - one always-warm gateway serves every app; there’s no per-app Azure resource and no per-app Entra registration to create.
The app is live, but locked: only the owner can open it, until your Ops team approves sharing with a group.
When the app shows corporate data, every query runs through the gateway, as the viewer. Nobody ever sees data their own permissions wouldn’t allow.

Rooted in your cloud. Deploy with a prompt.