Skip to content

Velonde documentation

Browser IDEs and app deployment for data-science teams

Velonde gives your team RStudio, VS Code and JupyterLab in the browser, with Claude Code and Posit Assistant already configured for the platform. When a project is ready, velonde deploy turns it into an app, report, API or pipeline that colleagues open with their company login. We can run Velonde on a dedicated machine for you, or you install it on your own server or Kubernetes cluster.

Who it is for

Velonde is for teams that work in R and Python and need to keep their code, data and apps inside the company: a data-science group in a mid-sized company, a research unit, or a consultancy working for several clients. Analysts and data scientists use it. Your IT department or we operate it. Nobody needs to know Docker, Kubernetes or YAML to publish an app for a colleague.

Features

Three IDEs, one home directory

RStudio Server, VS Code and JupyterLab start from the same page. Each user has a persistent home directory that all three IDEs share. You choose a size for each session and can run several sessions of one IDE side by side.

Coding agents that know the platform

Claude Code runs in every IDE, and RStudio also has Posit Assistant. Both receive the same instructions, rules and skills from the platform, so they know how projects are laid out, how a deploy works and how to fix a failed build. On request we set up another agent, or one that uses a model you host, and you can turn AI off entirely. Where your code and data go.

Deployment from the IDE

velonde deploy builds the project in the current directory, pushes the image and registers it. It handles Shiny for R and Python, Streamlit and Dash, Quarto, R Markdown and Jupyter notebooks, Plumber and FastAPI APIs including vetiver models, MCP servers in R and Python, pipelines of targets or of R and Python scripts, and reports that render on a schedule or after a pipeline, with parameters and a mail to their readers. velonde.yaml lists the groups that may open an app, and velonde rollback restores any of the last ten versions.

Your identity provider

Velonde works with Keycloak, Microsoft Entra ID, Okta, Auth0 and Cognito over OpenID Connect, with LDAP and Active Directory, and with SAML. The groups your provider sends decide who is an admin and who can see which app. With an OpenID Connect provider, APIs can accept its tokens too, so scripts and other services call them with a real identity.

Your packages and your network

The R and Python package stacks are each defined in a text file and pinned to a snapshot date. Private mirrors, company CA certificates and extra system packages are part of the configuration and reach every image. The image layer that contains them is built inside your network.

Versioned releases

Each release is a single versioned artifact. install.sh upgrade moves the workspace to a new version, and install.sh rollback returns to the previous one. Daily backups to an S3 bucket and monitoring with dashboards and alert mail are optional, one installer option each.

How a deploy works

  1. Open an IDE on the workspace and create a project: in RStudio with velonde::new_app("sales", template = "shiny") or File > New Project > New Directory > Velonde project, in VS Code or JupyterLab with velonde new sales --template shiny-python, or by asking the agent.
  2. Work as usual. renv or uv records the dependencies, and the agent knows the project layout and the workspace's commands.
  3. Run velonde deploy in the project directory. The workspace builds an image with your packages and their system libraries, pushes it to its registry and registers the app. The first build takes a few minutes. A redeploy with unchanged dependencies takes seconds.
  4. Colleagues open the app with their company login. velonde.yaml defines which groups have access.

This is the terminal output of a deploy, with the build log left out:

$ velonde deploy
velonde: packing /home/rstudio/sales as 'sales'
velonde: building sales as registry.velonde.acme.example/apps/sales:latest (shiny; mode: renv, Dockerfile: generated, renv.lock: yes)
velonde: system packages: libcurl4-openssl-dev libssl-dev
...
velonde: pushed registry.velonde.acme.example/apps/sales:latest (sha256:3f9c1a7e2b04)
velonde: app registered; the apps server restarts (about a minute)
velonde: https://apps.velonde.acme.example/app/sales

Hosted or self-hosted

Hosted by us

Each customer gets a dedicated server that is not shared with anyone else. We install, upgrade and monitor it, and your team logs in through your identity provider. The configuration belongs to you and moves with you if you later run Velonde yourself.

On your own infrastructure

The installer sets up a fresh Ubuntu server with one command. For an existing Kubernetes cluster, you run the same installer from a jump host. The platform images come from our registry with your account, and the layer with your packages and certificates is built on your side.

Built on open source

ShinyProxy and the ShinyProxy Operator run the sessions and the apps. k3s, Caddy, BuildKit, a container registry, Redis, Velero, Grafana, Loki and Prometheus provide the rest. No proprietary runtime sits between you and your code: velonde deploy produces a standard container image, and your projects are ordinary R or Python projects.

Try it

A trial installs with one command on any Ubuntu server and needs no DNS, only an account from us. Accounts and the trial explains what the account is for and what the trial includes. For an account, a hosted machine or a quote, write to stefan@linnerprogramming.com.