jupyter-kernel-gateway

What it is and what it does

Jupyter Kernel Gateway wraps Jupyter kernels in a web server, making them accessible over HTTP REST calls and WebSockets instead of the default ZeroMQ protocol. It runs kernels in its local process and filesystem, with no notebook editor—only kernel execution and management. The package supports two main modes: a WebSocket-based kernel protocol compatible with Jupyter Notebook's API, and a notebook-http mode that turns annotated notebook cells into HTTP endpoints, effectively turning notebooks into microservices.

The gateway is designed for scenarios where you want to attach notebooks to remote compute clusters, serve kernels to web clients that don't use the notebook interface, or deploy notebook logic as containerized microservices. It depends on jupyter-client, jupyter-core, jupyter-server, requests, tornado, and traitlets—all core Jupyter infrastructure—and can be scaled using container orchestration tools.

Use it for:

• Expose a Jupyter kernel over HTTP for a web dashboard or custom client application to execute code remotely
• Create a microservice from a notebook by annotating cells with HTTP verbs and serving them as REST endpoints
• Attach a local notebook environment to a remote compute cluster (e.g., Spark) without running a full notebook server
• Build a headless kernel pool for batch code execution via REST, without notebook editing overhead
• Integrate Jupyter kernel execution into a larger web application or cloud platform

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Jupyter Kernel Gateway is a web server that exposes Jupyter kernels over HTTP and WebSocket, enabling remote code execution and kernel management without a notebook interface.

Yes, if you need remote kernel execution over HTTP without a notebook interface. The package is stable, has low install friction, carries no known vulnerabilities, and is backed by the Jupyter organization. Install with caution if you require active feature development—the project is dormant and last released 885 days ago, so expect no near-term updates.

Install

jupyter-kernel-gateway on PyPI

pip install jupyter-kernel-gateway

uv add jupyter-kernel-gateway

poetry add jupyter-kernel-gateway

Installing jupyter-kernel-gateway

Before you install

Low friction install with six well-maintained runtime dependencies. The project is dormant (last release 885 days ago) but the repository remains active and supported by the Jupyter organization; suitable for stable production use where active development is not expected.

License in practice

Modified BSD License (3-Clause) is permissive and widely compatible. You may use, modify, and distribute this package freely in commercial and private projects provided you retain the license notice and disclaimer.

Quickstart

pip install jupyter-kernel-gateway

jupyter kernelgateway

# Or in Python:
from jupyter_kernel_gateway.notebook_http.handlers import NotebookHTTPHandler

Requires a local Jupyter kernel installation and Python 3.8 or later. The gateway spawns kernels in its local process space, so kernel dependencies must be available on the host.

Verify before relying

• Whether the package supports Python 3.12 or later (classifiers list up to 3.11)
• Current state of containerization and Kubernetes integration examples mentioned in the description

Package facts

Evidence: jupyter_kernel_gateway-3.0.1-py3-none-any.whl

Keywords: Cloud, Interactive, Interpreter, Kernel, Web

Tags

More WWW/HTTP packages