Practical overview
When it fits
Use envd when you want to define and build a container-based AI/ML development environment through a CLI, particularly for a Python workflow or one that includes Jupyter.
Who it is for
- AI/ML engineers who need isolated development environments with Python and system dependencies.
- Teams seeking reusable environment configuration instead of repeated Docker setup.
- Developers who want to include Jupyter in a container-based workflow.
Documented uses
- Declare development dependencies in a build.envd manifest.
- Build and enter an isolated development container with envd up.
- Enable Jupyter in an environment and retrieve its endpoint from the environment listing.
- Reuse build functions and import configuration from Git repositories.
Requirements & limits
The documented workflow is CLI-based and requires Docker 20.10.0 or later. The guide lists pip, uv, Homebrew, pipx, and an installation script as envd installation methods.
- The documented primary workflow requires Docker and command-line use.
- Environment setup is declared in build.envd and started with envd up.
- Jupyter is enabled through config.jupyter(), with the endpoint shown by envd envs ls.
- The guide describes BuildKit parallel builds and shared software caches, including PyPI and apt cache examples.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Set up a quick-start envd environment
Build and enter the documented Python-based container environment from the envd quick-start project.
Before you start
- Docker 20.10.0 or later.
- Install envd using one of the documented methods: pip, uv, Homebrew, pipx, or the installation script.
- Run envd bootstrap.
Steps
Clone the envd-quick-start repository.
Change into the cloned project directory.
Run envd up.
Expected outcome
The guide's example builds the environment and attaches the user to its container shell.
Guide limitations
- The example uses Python, Conda, and NumPy; the guide points to separate examples for R and Julia.
- The guide documents setup steps and an example outcome, not a compatibility guarantee for other configurations.
Enable Jupyter in an envd environment
Add Jupyter to the documented quick-start environment and retrieve its endpoint.
Before you start
- A bootstrapped envd installation and Docker 20.10.0 or later.
- The envd-quick-start project and its build.envd file.
Steps
Add config.jupyter() to the build() function in build.envd.
Start the environment with envd up --detach.
Run envd envs ls to view the Jupyter endpoint.
Expected outcome
The guide shows an environment listing with a Jupyter endpoint.
Guide limitations
- The documented endpoint is an example output; the guide does not specify a fixed endpoint value.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 12 October 2026
- Original website title
- envd | Development Environment for Machine Learning Engineers · Official homepage title · retained in its source language