Skip to main content

Python Ecosystems For Revit

Revit Python development has several runtime options. Choose based on the workflow you need: modern packages, direct .NET compatibility, visual programming, or team deployment.

Claimed hosts for RevitDevTool Python today: Revit and the AutoCAD-family.

Quick Choice​

NeedRecommended option
Modern Python packagesRevitDevTool CPython
VSCode breakpointsRevitDevTool CPython
Data analysis / ML / computational geometryRevitDevTool CPython
Existing IronPython or pyRevit scriptsRevitDevTool IronPython or pyRevit
Visual graph workflowDynamo Python
Team ribbon deploymentpyRevit
Simple no-package Revit API automationIronPython or compiled .NET

Runtime Comparison​

RuntimeBest forStrengthsTradeoffs
RevitDevTool CPythonpackage-heavy developmentmodern Python, PEP 723 dependencies, debugpy debugginguses PythonNet for .NET interop
RevitDevTool IronPythonIronPython scripts and existing pyRevit/RevitPythonShell-style codeOn Revit, execution and tests prefer the pyRevit engine (default IronPython 2.7.12); IronPython 3.4.2 fallback on Revit; bundled 3.4.2 on AutoCAD-familyno debugger attach; no modern CPython package ecosystem
pyRevit IronPythonribbon automation and team scriptsmature extension model, direct CLR styleolder Python ecosystem
pyRevit CPythonexperimentationCPython access inside pyRevit ecosystemmanual/runtime caveats vary by setup
Dynamo CPythongraph + Python workflowsvisual programming, bundled packagesless suited for code-first debugging
Compiled .NETproduction add-insstrong typing, host-native API model, attach debuggerslower iteration than scripts

RevitDevTool CPython​

Use this for new Python work when packages matter.

Good fit:

  • model data extraction;

  • dataframe analysis;

  • Excel/report generation;

  • geometry processing;

  • research scripts;

  • AI/ML experiments;

  • scripts that benefit from VSCode debugging.

Dependency example:


# /// script

# dependencies = [

#     "polars==1.38.1",

#     "numpy==2.4.2",

#     "openpyxl==3.1.5",

# ]

# ///

Interpreter owner​

PathHostsPython source
Pixi-ownedRevit, AutoCAD, Civil 3D, and other AutoCAD verticals (not Plant 3D)Pixi conda env at %APPDATA%\RevitDevTool\pixi-env (python = "3.14.*")
Host-attachedPlant 3D only — AutoCAD-family exceptionHost python3xx.dll via pythonnet — no second in-process interpreter

Package manager​

PathPrimaryFallback
Pixi-ownedPixi (conda-forge + PyPI, search-first per PEP 723)pip / pyRevit cengines
Plant 3D host-attacheduv sidecar (%APPDATA%\RevitDevTool\uv-env\{major.minor})pip matching host minor via pyRevit cengines

Pixi is the default package manager on Revit and AutoCAD-family. Plant 3D is the only AutoCAD-family vertical that embeds CPython before the add-in loads, so uv is used there and Pixi is not attempted.

For the full Plant 3D path, see the Plant 3D example and Modern Python Scripting.

IronPython​

Use IronPython when IronPython/.NET behavior or compatibility is more important than modern CPython packages.

Good fit:

  • existing pyRevit/RevitPythonShell-style scripts;

  • simple Revit API automation;

  • workflows that do not need NumPy, pandas, Polars, Shapely, or other CPython packages.

On Revit, IronPython scripts (*_ipy_script.py) and IronPython tests (test_*_ipy.py) both prefer the pyRevit engine when it is installed — pyRevit's current default is still IronPython 2.7.12, which matches most pyRevit tool work. If pyRevit is not installed, RevitDevTool falls back to IronPython 3.4.2.

On AutoCAD-family hosts, RevitDevTool uses bundled IronPython 3.4.2 (no pyRevit).

IronPython has no debugger attach in RevitDevTool. Use CPython when you need VSCode breakpoints.

The two filename conventions are for different purposes (execute vs pytest routing); they share the same Revit engine preference.

Dynamo Python​

Use Dynamo Python when the surrounding workflow is a Dynamo graph.

Good fit:

  • visual programming;

  • graph-controlled inputs/outputs;

  • Dynamo package ecosystem;

  • workflows used by designers who prefer node graphs.

pyRevit​

Use pyRevit when the main goal is distributing buttons and automations to users.

Good fit:

  • team ribbon tools;

  • established pyRevit extensions;

  • reusable buttons;

  • end-user automation workflows.

RevitDevTool can still be useful beside pyRevit for development, package experiments, and IronPython/CPython compatibility workflows. RevitDevTool also uses pyRevit cengines as a pip fallback when Pixi cannot run (or uv on Plant 3D).

Interface And .NET Interop Notes​

CPython uses PythonNet to talk to .NET APIs. This gives access to modern Python packages, but .NET interop is not identical to IronPython.

Watch for:

  • .NET interface implementation differences;

  • overload resolution;

  • extension methods;

  • out/ref style APIs;

  • dynamic object behavior inside the debugger.

For pure Revit API automation with heavy .NET interop and no Python packages, IronPython or compiled .NET may be simpler. For package-heavy work, CPython is usually the better choice.