Skip to content

Self-hosted · R and Python

Fast dashboards. Calm operations.

ShinyHub deploys and operates interactive R and Python applications with built-in authentication, scaling, hibernation, and observability—without turning your platform into another cloud console.

uv tool install shinyhub
shinyhub serve
ShinyHub dashboard showing running applications and their CPU and memory usage
One control plane for every dashboard
One binaryStart with SQLite and no external services.
Several runtimesRun R Shiny, Python Shiny, Dash, and Streamlit.
Production controlsAuthentication, isolation, scaling, and audit history.

From source to a dependable URL

  1. Validate locally

Run the production-shaped app route on your workstation and let doctor find bundle, runtime, and access problems before deployment.

  1. Preview the change

shinyhub plan shows the exact archive, manifest effects, permissions, and lifecycle changes without mutating the server.

  1. Deploy with progress

Follow dependency preparation, replica readiness, routing, and recovery through one stable CLI workflow.

  1. Operate by outcome

Keep viewer sessions responsive with hibernation, render pacing, worker isolation, metrics, and tracing.

shinyhub connect https://hub.example.com --name prod
shinyhub doctor ./my-app
shinyhub plan ./my-app
shinyhub deploy ./my-app --open

Start small, keep the path forward

Use the same Go binary for a local evaluation, a single Docker host, or a PostgreSQL-backed multi-node deployment. Runtime tiers let one control plane place applications locally, on remote Docker workers, or on AWS Fargate.

Choose an installation →

[app]
name = "Revenue forecast"
replicas = 2
max_sessions_per_replica = 10
render_seconds = 1.3

Built for everyone in the delivery chain

Developers get a direct deployment workflow. Operators get explicit state, bounded resources, and recovery paths. Viewers get the part that matters: a dashboard that opens quickly and stays dependable.

See a real ShinyHub instance

The public demo runs curated applications on an isolated app origin. Explore them anonymously, or sign in with GitHub for a read-only view of the actual control plane.

Open the demo guide