Python: Subdirectory `.venv` not detected by "Python: Select Interpreter"

[muzzlol]

Type: Bug

Behaviour

When creating a virtual environment (.venv) inside a subdirectory of the folder opened in VS Code, the interpreter does not show up in the Python: Select Interpreter dropdown. However, when the .venv is created in the root directory (where VS Code was opened), it is correctly detected.

---

Steps to reproduce:

1. Run the following commands in terminal:

   mkdir -p dir/subdir
   cd dir
   code .

2. In the VS Code workspace root (dir), run:

   python3 -m venv venv

3. Open Command Palette (Cmd+Shift+P) → Python: Select Interpreter
   ✅ The ./venv interpreter is listed.

4. Now instead of root, do this:

   cd subdir
   python3 -m venv venv

5. Go back to the Command Palette → Python: Select Interpreter
   ❌ The new subdir/venv interpreter is not listed, even though it exists and is valid.

Expected: All .venv folders within the workspace (even in subdirectories) should be detected or at least be selectable from the interface.

---

Diagnostic data

launch.json configuration

Output for Python in the Output panel

(no relevant logs - interpreter just not detected)

Output for Python Debugger in the Output panel

(no relevant logs - issue happens before debugging)

Extension version: 2025.4.0
VS Code version: Code 1.98.2 (Universal) (ddc367ed5c8936efe395cffeec279b04ffd7db78, 2025-03-12T13:32:45.399Z)
OS version: Darwin arm64 24.3.0
Modes:

• Python version (& distribution if applicable, e.g. Anaconda):
• Type of virtual environment used (e.g. conda, venv, virtualenv, etc.): Unknown
• Value of the python.languageServer setting: Default

User Settings

languageServer: "Pylance"

Installed Extensions

Extension Name	Extension Id	Version
ayu	tea	1.0.5
code-runner	for	0.12.2
codesnap	adp	1.3.4
copilot	Git	1.284.0
copilot-chat	Git	0.25.1
debugpy	ms-	2025.4.1
es7-react-js-snippets	dsz	4.4.3
figma-vscode-extension	fig	0.4.0
go	gol	0.46.1
js-debug	ms-	1.97.1
js-debug-companion	ms-	1.1.3
jupyter	ms-	2025.2.0
jupyter-keymap	ms-	1.1.2
jupyter-renderers	ms-	1.1.0
languague-renpy	cob	2.0.0
live-server	ms-	0.4.15
LiveServer	rit	5.7.9
makefile-tools	ms-	0.12.17
material-icon-theme	PKi	5.20.0
prettier-vscode	esb	11.0.0
python	ms-	2025.2.0
rainbow-csv	mec	3.18.0
remote-containers	ms-	0.401.0
TabOut	alb	0.2.2
vscode-docker	ms-	1.29.4
vscode-js-profile-table	ms-	1.0.10
vscode-jupyter-cell-tags	ms-	0.1.9
vscode-jupyter-slideshow	ms-	0.1.6
vscode-proto3	zxh	0.5.5
vscode-pull-request-github	Git	0.106.0
vscode-pylance	ms-	2025.3.2
vscode-tailwindcss	bra	0.14.9

System Info

Item	Value
CPUs	Apple M2 (8 x 2400)
GPU Status	2d_canvas: enabled canvas_oop_rasterization: enabled_on direct_rendering_display_compositor: disabled_off_ok gpu_compositing: enabled multiple_raster_threads: enabled_on opengl: enabled_on rasterization: enabled raw_draw: disabled_off_ok skia_graphite: disabled_off video_decode: enabled video_encode: enabled webgl: enabled webgl2: enabled webgpu: enabled webnn: disabled_off
Load (avg)	2, 3, 4
Memory (System)	8.00GB (0.14GB free)
Process Argv	. --crash-reporter-id 548eaa6f-0eef-4f3f-a3f7-ca08a4f49cc4
Screen Reader	no
VM	0%

A/B Experiments

vsliv368cf:30146710
vspor879:30202332
vspor708:30202333
vspor363:30204092
vscod805cf:30301675
binariesv615:30325510
py29gd2263:31024239
c4g48928:30535728
azure-dev_surveyone:30548225
a9j8j154:30646983
962ge761:30959799
h48ei257:31000450
pythontbext0:30879054
cppperfnew:31000557
dwnewjupytercf:31046870
nativerepl2:31139839
pythonrstrctxt:31112756
nativeloc2:31192216
iacca1:31171482
5fd0e150:31155592
dwcopilot:31170013
6074i472:31201624
dwoutputs:31242946
customenabled:31248079
9064b325:31222308
copilot_t_ci:31222730
f5992895:31259995
jda6j935:31233686
pythoneinst12:31262605
bgtreat:31262795
fh1c7952:31258891
6h6a5794:31261449
31787653:31262186
0339e702:31259951

[notuntoward]

Yes, exactly. It doesn't work for .py files nor for .ipynb files. I know my .venv is correctly configured because, from the command line, I can type:

uv run pytest

And then all the tests run. But they don't run from inside of vscode and neither do notebook files, where, in the upper right vscode corner, I see that the virtual environment is supposedly set to the right place:

[karthiknadig]

@notuntoward Can you file the notebook issue on https://github.com/microsoft/vscode-jupyter

I would like to better understand the intent of using a sub directory .venv. If the intent is that the sub directory should use a different virtual env than the parent directory. Or do you want to use different venv with the whole workspace, but they are created in sub directories?

Also when you run uv run pytest which directory are you running it from? From the sub directory or the parent directory?

[notuntoward]

OK, I'll file a bug on the jupyter repo.

I'm not familiar with the python design choices that led to the .venv folder, and its role in storing interpreters and packages. But it's the standard followed by everybody who uses virtual environments.

Somehow, vscode isn't following this ancient virtual environment convention.

I'm running uv run pytest in the "project root" directory, the place where the uv file pyproject.toml is located. The test directory just under it is where pytest looks for the test code. pytest runs any test*.py or *_test.py file recursively below the directory where it is started.

But this is a kind of side issue. vscode has the full path to python.exe yet it cannot find it.

[muzzlol]

I would like to better understand the intent of using a sub directory .venv. If the intent is that the sub directory should use a different virtual env than the parent directory. Or do you want to use different venv with the whole workspace, but they are created in sub directories?

Its both of those reasons yea. Im surprised this issue wasn't filed earlier. Maybe the convention is to have virtual envs in only the parent dir, but i've never followed it unless its a python only repo.

[karthiknadig]

@notuntoward Please share the logs from Output > Python. I don't think I understand where the .venv is placed. Would it be possible to share your project layout. What I want to see is how the .venv sits it relation to other parts. A Typical projects looks like this.

[muzzlol]

@karthiknadig as you can see workspace_venv is detected but the .venv inside subdir is not.

https://github.com/microsoft/vscode-python/blob/cf894bb91dd146c54dbcc6d2e446d267263b86fa/src/client/pythonEnvironments/creation/common/commonUtils.ts
needs to have a recursive function to check virtualenvs inside sub directories i guess

[karthiknadig]

@muzzlol Typically you should be using multi-root workspaces in these scenarios, where you have different folders with different frameworks. The reason we don't dig into every folder is for performance. But when you create a multi-root workspace, each root can have its own python so we do detect those.

For scenarios where the parent project is a different language or framework, but the subdir (or project) is python, then I recommend using muti-root workspace. https://code.visualstudio.com/docs/editing/workspaces/multi-root-workspaces

There is another scenario, "mono-repo" where all projects exist under one root. This is something that we don't yet have great support for, and we are working on that.