@@ -15,6 +15,7 @@
 import io
 import json
 import os
+import re
 import subprocess
 import unittest
 from pathlib import Path
@@ -285,6 +286,23 @@ class TestPublishedSets(unittest.TestCase):
     def test_published_cuda_versions(self):
         self.assertEqual(FILTER.SUPPORTED_CUDA_VERSIONS, ["cu130", "cu132", "cu134"])
 
+    def test_published_cuda_versions_are_documented(self):
+        # The install table on the getting started page is the only place a user is told
+        # these wheels exist. A train published without a row there is published to nobody,
+        # and a row left behind after a train stops being built sends people to an index
+        # with no wheel in it.
+        page = ROOT / "docs" / "source" / "getting-started.md"
+        text = page.read_text(encoding="utf-8")
+        heading = "\n## Installation\n"
+        self.assertIn(heading, text, f"no Installation section in {page.name}")
+        section = text.split(heading, 1)[1].split("\n## ", 1)[0]
+        documented = set(re.findall(r"cu\d+", section))
+        self.assertEqual(
+            documented,
+            set(FILTER.SUPPORTED_CUDA_VERSIONS),
+            f"the install table in {page.name} lists {sorted(documented)}",
+        )
+
     def test_published_cuda_versions_are_supported_by_the_installer(self):
         supported = {
             f"cu{major}{minor}"
 
@@ -15,24 +15,44 @@ The following are required to install the ExecuTorch host libraries, needed to e
 - On Windows, Visual Studio 2022 or later.
 
 ## Installation
-To use ExecuTorch, you will need to install both the Python package and the appropriate platform-specific runtime libraries. Pip is the recommended way to install the ExecuTorch python package.
+To use ExecuTorch, you will need to install both the Python package and the appropriate platform-specific runtime libraries. Pip is the recommended way to install the ExecuTorch python package. Consider installing it within a virtual environment, such as one provided by [conda](https://docs.conda.io/projects/conda/en/latest/user-guide/getting-started.html#creating-environments) or [venv](https://packaging.python.org/en/latest/guides/installing-using-pip-and-virtual-environments/#create-and-use-virtual-environments).
 
-This package includes the dependencies needed to export a PyTorch model, as well as Python runtime bindings for model testing and evaluation. Consider installing ExecuTorch within a virtual environment, such as one provided by [conda](https://docs.conda.io/projects/conda/en/latest/user-guide/getting-started.html#creating-environments) or [venv](https://packaging.python.org/en/latest/guides/installing-using-pip-and-virtual-environments/#create-and-use-virtual-environments).
+Install PyTorch in the same command. The ExecuTorch package does not declare it
+as a dependency, because the build you need depends on your hardware, so pip
+cannot choose one for you. Installing ExecuTorch on its own gives an environment
+where exporting a model stops with `No module named 'torch'`.
 
-```
-pip install executorch
-```
-
-To get the latest features before they reach a stable release, install a nightly
-build instead. Nightly wheels are built from the `main` branch every day, so a
-change that has landed but is not yet in a stable release is available there first.
+Both packages come from the same package index. Find your machine in the table
+below and put the name from it in place of `<variant>`:
 
 ```
-pip install --upgrade --pre executorch torch --extra-index-url https://download.pytorch.org/whl/nightly/cpu
+pip install executorch torch \
+  --index-url https://download.pytorch.org/whl/<variant> \
+  --extra-index-url https://pypi.org/simple
 ```
 
-`torch` is explicit because nightly ExecuTorch wheels do not declare it as a
-dependency.
+| Machine you export on | Variant |
+| --- | --- |
+| CPU only | `cpu` |
+| NVIDIA GPU, CUDA 13.0 | `cu130` |
+| NVIDIA GPU, CUDA 13.2 | `cu132` |
+| NVIDIA GPU, CUDA 13.4 | `cu134` |
+
+The CUDA packages are built for Linux, on x86_64 and ARM64. Use the `cpu`
+variant on macOS and on Windows. If your CUDA version is not in the table,
+choose the closest lower one with the same major version, because CUDA works
+across minor versions but not across major ones. There is no package for CUDA
+12, so on CUDA 12 use the `cpu` variant or build from source.
+
+To get a change that has landed on the `main` branch but is not in a release
+yet, use a nightly build. These are rebuilt every day. Put `nightly/` in front
+of the variant name, for example `nightly/cu130`, and add `--pre` to the
+command, otherwise pip skips development versions. Nightly builds cover the same
+variants. CUDA 13.4 is the newest, and until the next release it is in nightly
+builds only.
+
+The second index is needed because a bare `--index-url` replaces PyPI instead of
+adding to it, and some dependencies are published only on PyPI.
 
 To build the framework from source, see [Building From Source](using-executorch-building-from-source.md). Backend delegates may require additional dependencies. See the appropriate backend documentation for more information.