Stop editing your test files just to run a subset — use pytest-web instead
You've been there. You have a test file with 40 parametrized cases. You need to run 3 of them. So you open the file, comment out the ones you don't want, run pytest, then uncomment everything again. Or you try to remember the exact -k expression. Or you copy-paste a node ID from the terminal and hope you got it right.
There's a better way.
What is pytest-web?
pytest-web is a local web UI that sits on top of your existing pytest setup. You run one command, a browser tab opens, and from there you can see all your tests, pick exactly which ones to run, watch them pass or fail in real time, and re-run failures with a single click.
It doesn't replace pytest. It doesn't touch your conftest.py or your fixtures. It just wraps the same pytest command you already use and gives you a visual layer on top.
Getting started takes 30 seconds
pip install pytest-web
cd your-project
pytest-web
That's it. A browser tab opens at http://127.0.0.1:8000. Your project's pytest.ini, fixtures, and plugins are all picked up automatically — nothing to configure.
The parametrize problem, solved
Here's the situation that motivated this tool. Say you have this in your test file:
@pytest.mark.parametrize("browser,env", [
("chrome", "staging"),
("chrome", "production"),
("firefox", "staging"),
("firefox", "production"),
("safari", "staging"),
("safari", "production"),
])
def test_checkout_flow(browser, env):
...
Six variants. You've just pushed a fix and you only want to run the two production ones to verify.
Before pytest-web: comment out 4 parametrize entries, run pytest, uncomment them. Or write -k "production" and hope that expression doesn't accidentally catch something else.
With pytest-web: click Fetch Tests. Every variant appears as its own row:
✅ test_checkout_flow[chrome-staging]
✅ test_checkout_flow[chrome-production]
✅ test_checkout_flow[firefox-staging]
✅ test_checkout_flow[firefox-production]
✅ test_checkout_flow[safari-staging]
✅ test_checkout_flow[safari-production]
Uncheck the four staging ones. Click ▶ Run Selected. Done. Your file is untouched.
Watch results in real time
While tests run, each row shows a pulsing dot. As each test finishes it turns:
- 🟢 Green — passed
- 🔴 Red — failed
- 🟡 Yellow — skipped
You don't wait for the whole suite to finish. You see results as they happen. If something fails early you already know while the rest are still running.
Re-run only the failures
This is the feature people end up using the most.
After a run, click the failed counter at the top. The view filters to show only failed tests. Click ▶ Run Selected. Only those tests run — nothing else.
No -k expression. No copy-pasting node IDs. No editing any file. Just two clicks.
This loop — run → filter to failed → fix → re-run failures — is where pytest-web saves the most time day to day.
Run tests in parallel
Set the Workers field to any number and pytest-web runs your suite with pytest-xdist automatically. No config changes needed:
- Workers:
1→ normal sequential run - Workers:
4→ 4 parallel processes
pytest-xdist is installed as a dependency, so it's already there. pytest-web handles all the complexity of deduplicating events from parallel workers so the UI stays correct.
Inject env vars without touching your shell
Need to run tests against a different environment? Use the Env Vars panel:
| Name | Value |
|---|---|
ENV | staging |
API_URL | https://api.staging.example.com |
Variables are saved to localStorage and restored on page reload. Type NAME=value in the Name field and it auto-splits — no need to fill both columns separately.
Allure reports in one click
If you use Allure for test reporting, pytest-web has you covered. After your tests run:
- Click Open Report in the Allure bar at the bottom
- pytest-web generates the report and opens it in your browser automatically
- Every time you regenerate, history is carried forward — so trend graphs and retry counts work out of the box
No commands to remember. No manual allure generate && allure open. Just one click.
Everything you need, nothing you don't
Here's the full feature list — and notice how short it is:
| Feature | How it works |
|---|---|
| Fetch tests | Click "Fetch Tests" or press Enter |
| Select/deselect | Checkboxes per test or per file |
| Filter by name | Type in the search box |
| Filter by outcome | Click passed / failed / skipped counter |
| Parallel workers | Set the number in the Workers field |
| Custom pytest args | Type in the args bar or use the param builder dropdown |
| Env var injection | Add rows in the Env Vars panel |
| Live output log | Collapsible panel at the bottom, opens on errors |
| Copy command | Click ⎘ to copy the exact pytest command |
| Allure report | Click "Open Report" in the Allure bar |
| Dark mode | Theme switcher in the top right |
Try it now
pip install pytest-web
cd your-project
pytest-web
If you run into anything or have a feature idea, open an issue on GitHub. The project is MIT licensed and contributions are very welcome.