Library / SDK
panda3d/panda3d avatar
panda3d/panda3d

The documentation link points at 1.11 while the SDK it tells you to install is 1.10.16

Powerful, mature open-source cross-platform game engine for Python and C++, developed by Disney and CMU

5,272 stars902 forksC++NOASSERTION

At a glance

What is it?
Panda3D is a long-lived engine and framework for 3D rendering and game development with bindings for Python and C++, and its readme is a build manual: four platforms, one Python build script, and a set of optional dependencies you either take wholesale or pick from. The version numbers on the page do not all agree with each other.
Who is it for?
Panda3D suits a project that wants scripting in Python without giving up a C++ engine underneath, and that values a permissive licence and a build you can take apart. Three things to check before you commit.
Can I use it commercially?
Check first. The repository uses a licence we do not classify automatically, so read its LICENSE file before any commercial use.
Is it still maintained?
Yes. The repository last received commits 110 days ago.
What is it written in?
Mainly C++, according to GitHub's language statistics.

Answers come from the project's GitHub data, last synced on November 15, 2026, and from our analysis. They are not legal advice.

From daily snapshots

Star history

  • panda3d/panda3d5,272 ★ · +45
Star history
Repository2026-09-232026-11-15
panda3d/panda3d5.2k5.3k

45 Stars in the last 53 days, measured from the daily GitHub snapshot. Compare with other projects

Editorial analysis

The documentation link points at a newer line than the download

The readme is careful about versions in some places and not in others. The SDK download page it links is for 1.10.16, and the two archives of prebuilt dependencies it gives you for Windows and one for macOS are all named with that same version. The release list agrees: the three most recent are 1.10.16, 1.10.15 and 1.10.14, titled with a product prefix rather than a bare version number. But the page it sends you to for an overview of the various third-party dependencies lives under a documentation path for 1.11. That is the inconsistency worth knowing about, because it is exactly the page you need when the build complains that it cannot find something: the list you are reading may describe a release you have not installed. It is a small thing, and it is the kind of small thing that costs an afternoon when a dependency name in the newer documentation does not match the one your build is looking for.

Two build systems are in the tree and only one is documented

Every build instruction on the page runs the same Python script, invoked through whichever interpreter you want the engine built against. That script lives in its own directory beside the command line tools and beside a second installation script that copies files directly when your distribution has no package format. But the repository root also carries a CMake build file, a CMake directory, a contrib directory, a direct/ directory, and a dtool/ directory. So there are two build systems here, and every build command on the page uses the makepanda script. That is not necessarily a problem: a project of this age often keeps a native build for embedding and a script for standalone installers. But anyone whose toolchain expects the native build will not find it in these build instructions, and anyone who assumes the script is the only option will not know the alternative exists.

Each platform produces a different artifact

The four platform sections converge on one script and diverge on what you get back. On Windows a successful build produces an executable installer. On Linux the installer flag produces a Debian or RPM package depending on your distribution, and if neither format suits you there is the direct-copy script, after which you may have to refresh the library cache yourself. On macOS the same flags produce a disk image containing an installer package. And on Linux there is a third option the other platforms do not mention, a flag that produces a Python wheel instead, installable with the package installer. That last one is the least discussed and the most convenient if you already ship Python, and it is the reason the page can offer both a binary SDK and a package-manager install without them being rivals. The common flag across all of them turns on every feature for which the script can detect a prerequisite, which is also why you will see warnings about missing packages on a first run.

Windows support for an old operating system costs a different SDK and a different flag

The Windows section is the longest and the most conditional. You need a Microsoft compiler, of which three versions are named, and the Windows SDK, and then a conditional clause: if you intend to target Windows Vista you also need the Windows 8.1 SDK, and you must remove the SDK option from the build command. The version mapping is given inline as a lookup table, with 2022 mapping to one number, 2019 to the one before, 2017 to the one before that, and 2015 to a round number. That last entry is worth noticing, because the compiler list at the top of the section does not include 2015 while the mapping below it does. The rest of the section is dependencies: two archives of prebuilt third-party libraries, one per architecture, or instructions for building them from source in a separate repository. So the cost of that extra platform is a different SDK and a different flag.

Eigen support is documented as a build that takes hours

There is a short note in the Windows section, and it is the most honest sentence in the build documentation. The command by default turns off a linear algebra library, and the note says you may remove that flag to improve runtime performance, but that the build will then take hours to complete, because the library is heavily template-based and the compiler does not do well in that situation. So the default is a fast build with slower linear algebra, and the alternative is a slow build with faster linear algebra. The note does not say how many hours, and does not offer a middle path such as precompiled headers. For a project whose selling point is that it works on four platforms, this is the kind of note that tells you the build is not as turnkey as the section heading suggests.

The Linux dependency list is one very long apt command

After the build command, the page warns that you will probably see warnings about missing dependency packages and tells you to decide which you want. Then it offers a single apt command with roughly twenty-five packages on one line, covering image libraries, audio libraries, windowing, input, font rendering, a physics engine, an importer for scene formats, open image formats, a compression library, and an NVIDIA toolkit. It also links a manual page for the full set of dependencies and their licences, which is the page I would read before shipping anything commercially, since it is where the bundled third-party licences are enumerated. FreeBSD gets the same treatment with a package manager line containing a partly different list, including a couple of packages the Debian list does not mention.

Development builds come from an archive index rather than the package index

The stable install is one line:

bash
pip install panda3d

The development install is two options. The first adds a pre-release flag and points the installer at a separate index hosted on the project's own archive domain, which is how a nightly build of a compiled extension gets distributed without waiting for the main package index. The page attaches a troubleshooting note to it: if the command fails, make sure your installer is up to date, which is a reasonable thing to say about a command that uses a flag not every version understands. The second option is the full SDK with all the tools, from a page that says the builds there are kept in step with the latest state of the repository automatically. That last phrasing is the one to notice, because it means the development SDK moves under you, which is fine for testing and not fine for a reproducible build.

The licence is described twice and the metadata records neither

Two paragraphs say two different things about the same file. The first says the engine is open source and free for any purpose including commercial ventures, thanks to what it calls a liberal licence, and points at a licence page on the project site. The second says the engine is licensed under the Modified BSD licence and to see the licence file for details. Those are compatible statements, but the first is written as marketing and the second as a pointer, and the repository's licence metadata reads NOASSERTION, so a tool reading metadata gets no licence identifier. Around the licence sit the other obligations of a mature project: a backers file, a documentation page enumerating the third-party licences for the bundled dependencies, configuration for two Python static analysers, a requirements file for the test dependencies, and sample projects that each demonstrate one thing, from physics and culling to mouse input, networking, glow and distortion filters.

Editorial conclusion

Panda3D suits a project that wants scripting in Python without giving up a C++ engine underneath, and that values a permissive licence and a build you can take apart. Three things to check before you commit. The readme points at documentation for a newer line than the SDK it tells you to download, so read the version-specific notes for the line you actually install. The readme's build commands all use makepanda, while the repository also carries a CMakeLists.txt that your toolchain may prefer. And the dependency story is all-or-nothing by default: the everything flag turns on every feature whose prerequisites it can detect, so most first builds produce warnings about missing packages that you then have to decide about one by one.

Frequently asked questions

what is panda3d

A game engine and a framework for 3D rendering and game development, with bindings for Python and C++ programs, open source and free for any purpose including commercial use. The readme points at a gallery of examples, a feature list, documentation resources, and a community forum for help.

how to install panda3d using pip

One command installs the stable release from the package index. For the latest development build you add a pre-release flag and point the installer at a separate archive index hosted on the project's own domain, and the page notes that if that command fails you should update your installer first. A separate page offers full SDK builds with all the tools, which it says are kept in step with the repository automatically.

is panda3d open source

Yes. The readme says it is open source and free for any purpose, including commercial ventures, and separately states that it is licensed under the Modified BSD licence, pointing at the licence file. The repository's licence metadata reads NOASSERTION, and the bundled third-party dependencies have their own licences enumerated on a documentation page.

How old is Panda3D?

No founding date is given. The current SDK is 1.10.16, the three most recent releases are 1.10.16 (2025-12-25), 1.10.15 (2024-11-08) and 1.10.14 (2024-01-10), and the documentation link it gives for third-party dependencies points at a 1.11 path. The repository description credits its development to Disney and CMU.

how to install panda3d on linux

Besides `pip install panda3d`, you can build it and then install the package. The build script with the everything flag turns on every feature whose prerequisites it can detect, so expect warnings about missing packages and a single apt command covering around twenty-five of them. You then install the resulting Debian or RPM package, or, if your distribution supports neither, run the script that copies files into place and refresh the library cache yourself. There is also a flag that produces a Python wheel instead, which the page mentions only on this platform.

Official sources

  1. Issues
  2. panda3d/panda3d on GitHub
  3. Project website
  4. README
  5. Releases
Add this badge to your README

If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.

Add this badge to your README

markdown
[![Hysen Labs](https://hysenlabs.com/badge/panda3d-panda3d.svg)](https://hysenlabs.com/projects/panda3d-panda3d)

Star history chart for your README

markdown
[![Star history](https://hysenlabs.com/chart/panda3d-panda3d.svg)](https://hysenlabs.com/projects/panda3d-panda3d)

Was this useful?