diff --git a/website/src/widgets/quickstart-install.js b/website/src/widgets/quickstart-install.js
index 7e3bed136..df8d971f0 100644
--- a/website/src/widgets/quickstart-install.js
+++ b/website/src/widgets/quickstart-install.js
@@ -19,54 +19,6 @@ const CUDA = {
     '11.0': 'cuda110',
 }
 const LANG_EXTRAS = ['ja'] // only for languages with models
-const DATA = [
-    {
-        id: 'os',
-        title: 'Operating system',
-        options: [
-            { id: 'mac', title: 'macOS / OSX', checked: true },
-            { id: 'windows', title: 'Windows' },
-            { id: 'linux', title: 'Linux' },
-        ],
-    },
-    {
-        id: 'package',
-        title: 'Package manager',
-        options: [
-            { id: 'pip', title: 'pip', checked: true },
-            { id: 'conda', title: 'conda' },
-            { id: 'source', title: 'from source' },
-        ],
-    },
-    {
-        id: 'hardware',
-        title: 'Hardware',
-        options: [
-            { id: 'cpu', title: 'CPU', checked: DEFAULT_HARDWARE === 'cpu' },
-            { id: 'gpu', title: 'GPU', checked: DEFAULT_HARDWARE == 'gpu' },
-        ],
-        dropdown: Object.keys(CUDA).map(id => ({ id: CUDA[id], title: `CUDA ${id}` })),
-        defaultValue: DEFAULT_CUDA,
-    },
-    {
-        id: 'config',
-        title: 'Configuration',
-        multiple: true,
-        options: [
-            {
-                id: 'venv',
-                title: 'virtual env',
-                help: 'Use a virtual environment and install spaCy into a user directory',
-            },
-            {
-                id: 'train',
-                title: 'train models',
-                help:
-                    'Check this if you plan to train your own models with spaCy to install extra dependencies and data resources',
-            },
-        ],
-    },
-]
 
 const QuickstartInstall = ({ id, title }) => {
     const [train, setTrain] = useState(false)
@@ -100,7 +52,56 @@ const QuickstartInstall = ({ id, title }) => {
                 const pkg = nightly ? 'spacy-nightly' : 'spacy'
                 const models = languages.filter(({ models }) => models !== null)
                 const data = [
-                    ...DATA,
+                    {
+                        id: 'os',
+                        title: 'Operating system',
+                        options: [
+                            { id: 'mac', title: 'macOS / OSX', checked: true },
+                            { id: 'windows', title: 'Windows' },
+                            { id: 'linux', title: 'Linux' },
+                        ],
+                    },
+                    {
+                        id: 'package',
+                        title: 'Package manager',
+                        options: [
+                            { id: 'pip', title: 'pip', checked: true },
+                            !nightly ? { id: 'conda', title: 'conda' } : null,
+                            { id: 'source', title: 'from source' },
+                        ].filter(o => o),
+                    },
+                    {
+                        id: 'hardware',
+                        title: 'Hardware',
+                        options: [
+                            { id: 'cpu', title: 'CPU', checked: DEFAULT_HARDWARE === 'cpu' },
+                            { id: 'gpu', title: 'GPU', checked: DEFAULT_HARDWARE == 'gpu' },
+                        ],
+                        dropdown: Object.keys(CUDA).map(id => ({
+                            id: CUDA[id],
+                            title: `CUDA ${id}`,
+                        })),
+                        defaultValue: DEFAULT_CUDA,
+                    },
+                    {
+                        id: 'config',
+                        title: 'Configuration',
+                        multiple: true,
+                        options: [
+                            {
+                                id: 'venv',
+                                title: 'virtual env',
+                                help:
+                                    'Use a virtual environment and install spaCy into a user directory',
+                            },
+                            {
+                                id: 'train',
+                                title: 'train models',
+                                help:
+                                    'Check this if you plan to train your own models with spaCy to install extra dependencies and data resources',
+                            },
+                        ],
+                    },
                     {
                         id: 'models',
                         title: 'Trained pipelines',
@@ -142,11 +143,6 @@ const QuickstartInstall = ({ id, title }) => {
                         setters={setters}
                         showDropdown={showDropdown}
                     >
-                        {nightly && (
-                            <QS package="conda" comment prompt={false}>
-                                # 🚨 Nightly releases are currently only available via pip
-                            </QS>
-                        )}
                         <QS config="venv">python -m venv .env</QS>
                         <QS config="venv" os="mac">
                             source .env/bin/activate