Choose the right Deephaven Python packages to install

This cheat sheet describes the Python packages on PyPI that you use with Deephaven, when to install each one, and which legacy or unsupported packages a PyPI search also returns. Start with the table below, then see the entry for each package.

Which package to install

To...Install
Write queries in a Deephaven server that runs from a Python Docker image, such as ghcr.io/deephaven/serverNothing. Every Deephaven Python image includes deephaven-core. Only ghcr.io/deephaven/server also includes the Deephaven Express, deephaven.ui, and Matplotlib plugins.
Write queries in a Deephaven server that runs from the production applicationdeephaven-core[autocomplete], installed into the server's virtual environment
Run a Deephaven server from Python, without Dockerdeephaven-server
Connect to a running Deephaven server from another Python processpydeephaven
Subscribe to ticking tables from a Python clientpydeephaven-ticking
Plot Deephaven tablesdeephaven-plugin-plotly-express (preinstalled in ghcr.io/deephaven/server)
Build real-time dashboardsdeephaven-plugin-ui (preinstalled in ghcr.io/deephaven/server)
Display Matplotlib and Seaborn plotsdeephaven-plugin-matplotlib (preinstalled in ghcr.io/deephaven/server). For Seaborn, install deephaven-plugin-matplotlib[seaborn].
Create TradingView Lightweight Chartsdeephaven-plugin-tradingview-lightweight
Add color themes to the Deephaven Web UIdeephaven-plugin-theme-pack
Write a Deephaven server-side plugindeephaven-plugin
Use local Python workspaces from the Deephaven VS Code extensiondeephaven-plugin-python-remote-file-source
Use Deephaven in Jupyterdeephaven-ipywidgets[server] to run the server in the notebook, or deephaven-ipywidgets[client] to connect to a running server
Display Deephaven tables in a Streamlit appstreamlit-deephaven
Trade through Interactive Brokersdeephaven-ib
Connect AI tools to Deephavendeephaven-mcp

Server and engine

deephaven-core

deephaven-core is the Deephaven Python Integration Package. It lets Python code access data, run queries, and execute scripts directly inside a Deephaven server, next to the data, on static and real-time tables alike. Python code imports it as deephaven. A legacy Enterprise package on PyPI, deephaven, also installs a module called deephaven, but that module is unrelated.

deephaven-core runs only inside a Deephaven server, and the Deephaven Web UI's Python console has it available at runtime. The Deephaven Python Docker images, such as ghcr.io/deephaven/server, include it. For a server run from the production application, install it into the server's virtual environment.

When you install it yourself, use deephaven-core[autocomplete]. The [autocomplete] suffix names an extra. An extra is an optional set of dependencies that pip installs only when you name the extra in brackets after the package name. This extra installs the libraries that provide code completion in the Deephaven Web UI.

deephaven-server

deephaven-server is a Python-native way to launch Deephaven Community Core. It embeds the server in a Python process, so you can install it with pip and run it without Docker. Installing it also installs deephaven-core[autocomplete] and jpy. You need Java 17 or later. See Install and run Deephaven with pip for the full prerequisites.

jpy

jpy is the bidirectional Python-Java bridge that deephaven-core and deephaven-server depend on. The jpy consortium maintains it, and installing either package installs jpy as a dependency. See Use jpy.

Clients

pydeephaven

pydeephaven is the Deephaven Python Client, a client API that lets Python applications access a Deephaven server remotely. See the Python client quickstart.

pydeephaven-ticking

pydeephaven-ticking extends pydeephaven so a client can subscribe to ticking tables and receive the added, removed, and modified rows of each update. It installs pydeephaven as a dependency. PyPI has prebuilt wheels for Linux x86_64 only. For other platforms, see the version support matrix. For usage, see Subscribe to ticking tables from a Python client.

Plugins and plugin development

Install a plugin into the Python environment the server runs in, then restart the server, because the server loads plugins only when it starts. In Docker, add the pip install line to a custom image rather than running it in a container. See Install and use plugins.

deephaven-plugin-plotly-express

deephaven-plugin-plotly-express is Deephaven Express, a Deephaven charting plugin built on Plotly Express. It is the recommended way to plot Deephaven tables. The ghcr.io/deephaven/server Docker image includes it.

deephaven-plugin-ui

deephaven-plugin-ui is the plugin for building real-time dashboards with deephaven.ui. The ghcr.io/deephaven/server Docker image includes it.

deephaven-plugin-matplotlib

deephaven-plugin-matplotlib is the Deephaven plugin for Matplotlib. It displays Matplotlib and Seaborn plots in Deephaven. The ghcr.io/deephaven/server Docker image includes it but not Seaborn. To install Seaborn as well, install the deephaven-plugin-matplotlib[seaborn] extra. See Matplotlib and Seaborn.

deephaven-plugin-tradingview-lightweight

deephaven-plugin-tradingview-lightweight is a Deephaven plugin for creating TradingView Lightweight Charts from Python. The ghcr.io/deephaven/server Docker image does not include it, so install it as described at the top of this section.

deephaven-plugin-theme-pack

deephaven-plugin-theme-pack adds color themes to the Deephaven Web UI. The ghcr.io/deephaven/server Docker image does not include it, so install it as described at the top of this section. See Install and use plugins and Create custom themes.

deephaven-plugin

deephaven-plugin is the Python interface for writing Deephaven server-side plugins. deephaven-core depends on it, so installing deephaven-core also installs deephaven-plugin. Install it directly only when developing a plugin. See Create your own plugin.

deephaven-plugin-python-remote-file-source

deephaven-plugin-python-remote-file-source is a bidirectional plugin that lets Python code import modules from a remote file source. The Deephaven VS Code extension uses it to enable local Python workspaces. The ghcr.io/deephaven/server Docker image does not include it, so install it as described at the top of this section.

deephaven-plugin-utilities

deephaven-plugin-utilities is a helper library that Deephaven plugins use at runtime. It is not a plugin, and you do not install it directly. Installing one of the plugins above, such as deephaven-plugin-plotly-express or deephaven-plugin-ui, installs it as a dependency.

deephaven-plugin-packaging

deephaven-plugin-packaging is a helper library that the Deephaven plugin build process uses to create plugin wheels. It is not a plugin, and you do not install it directly. Installing a plugin from a prebuilt wheel does not install it.

Notebooks, apps, and integrations

deephaven-ipywidgets

deephaven-ipywidgets displays Deephaven tables, figures, and other Deephaven objects as interactive widgets in Jupyter notebooks. The deephaven-ipywidgets[server] extra also installs deephaven-server, and the deephaven-ipywidgets[client] extra also installs pydeephaven. See Use Deephaven in Jupyter.

streamlit-deephaven

streamlit-deephaven displays Deephaven tables and other Deephaven widgets in a Streamlit app. It runs the Deephaven server through deephaven-server, which it installs as a dependency.

deephaven-ib

deephaven-ib is an Interactive Brokers integration for Deephaven. It runs on deephaven-server, which it installs as a dependency. See the deephaven-ib README for setup and usage.

deephaven-mcp

deephaven-mcp provides Model Context Protocol (MCP) servers that connect AI tools to Deephaven. See Deephaven MCP.

Legacy and unsupported packages

These packages appear in a PyPI search for Deephaven, but Deephaven Community Core does not use them. A PyPI search also returns experimental plugins published by individual developers, which Deephaven does not document or support.

deephaven-jpy

deephaven-jpy is a legacy Deephaven fork of jpy, last released in 2022. deephaven-core and deephaven-server depend on jpy instead and do not use deephaven-jpy.

deephaven-plugin-plotly

deephaven-plugin-plotly is a legacy Deephaven plugin for displaying Plotly figures. Deephaven no longer supports it. To display a Plotly figure, use deephaven-plugin-plotly-express, which also renders raw Plotly figures.

deephaven-plugin-plotly-js

deephaven-plugin-plotly-js packaged the JavaScript side of the legacy deephaven-plugin-plotly. PyPI has only development pre-releases of it. Use deephaven-plugin-plotly-express instead.

deephaven-plugin-json

deephaven-plugin-json is a Deephaven plugin for JSON that is available on PyPI only as development pre-releases. It has no stable release to install.

deephaven-example-app

deephaven-example-app is an example Deephaven application built on deephaven-server that ingests a real-time data feed, transforms it, and displays it in the Web UI. It was last released in 2022.

deephaven

deephaven is a legacy Python package for Deephaven Enterprise, last released in January 2023. It is not part of Deephaven Community Core. It installs a module that is also imported as deephaven but is unrelated to the deephaven module that Community queries use, which comes from deephaven-core. For Deephaven Enterprise, use the deephaven-coreplus-client Python client.