# What will i do?

**URL:** <https://discuss.roboflow.com/t/what-will-i-do/11722>\
**Category:** 🤝  Community Help\
**Tags:** export\
**Created:** [January 21, 2026, 8:10am UTC](https://discuss.roboflow.com/t/what-will-i-do/11722 "2026-01-21T08:10:06Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![Даниил](https://yyz1.discourse-cdn.com/flex029/user_avatar/discuss.roboflow.com/%D0%B4%D0%B0%D0%BD%D0%B8%D0%B8%D0%BB/32/8959_2.png) [@Даниил](https://discuss.roboflow.com/u/%D0%94%D0%B0%D0%BD%D0%B8%D0%B8%D0%BB)\
**Post date:** [January 21, 2026, 8:10am UTC](https://discuss.roboflow.com/t/what-will-i-do/11722/1 "2026-01-21T08:10:06Z")

</div>

I just want the AI to see my screen and highlight things from the dataset on it. What should I do once the neural network is trained?

_Describe your question/issue here! (delete this when you post)_

- **Project Type:**
- **Operating System & Browser:**
- **Project Universe Link or Workspace/Project ID:**
- **Do you grant Roboflow Support permission to access your Workspace for troubleshooting? (Yes/No):**

---

<div class="post-metadata">

**Author:** ![Automatez](https://yyz1.discourse-cdn.com/flex029/user_avatar/discuss.roboflow.com/automatez/32/8971_2.png) [@Automatez](https://discuss.roboflow.com/u/Automatez)\
**Post date:** [January 21, 2026, 11:11pm UTC](https://discuss.roboflow.com/t/what-will-i-do/11722/2 "2026-01-21T23:11:29Z")

</div>

Got an answer from AI you could try - you use OBS Studio to stream your screen as a virtual webcam and run your model against that. Code (untested) is below.

```auto
import cv2
import base64
import numpy as np
import requests

ROBOFLOW_API_KEY = "YOUR_KEY"
ROBOFLOW_MODEL = "your-model-id"
ROBOFLOW_SIZE = 416

upload_url = "".join([
    "https://detect.roboflow.com/",
    ROBOFLOW_MODEL,
    "?access_token=",
    ROBOFLOW_API_KEY,
    "&format=image",
    "&stroke=5"
])

# Use OBS virtual camera instead of default webcam
video = cv2.VideoCapture(1) # 0 = default cam, 1 = OBS virtual cam on many systems

def infer():
    ret, img = video.read()
    height, width, _ = img.shape
    scale = ROBOFLOW_SIZE / max(height, width)
    img = cv2.resize(img, (round(scale * width), round(scale * height)))

    retval, buffer = cv2.imencode('.jpg', img)
    img_str = base64.b64encode(buffer)

    resp = requests.post(upload_url, data=img_str, headers={
        "Content-Type": "application/x-www-form-urlencoded"
    }, stream=True).raw

    image = np.asarray(bytearray(resp.read()), dtype="uint8")
    image = cv2.imdecode(image, cv2.IMREAD_COLOR)
    return image

while True:
    if cv2.waitKey(1) == ord('q'):
        break
    image = infer()
    cv2.imshow('image', image)

video.release()
cv2.destroyAllWindows()

```

---

<div class="post-metadata">

**Author:** ![system](https://canada1.discourse-cdn.com/flex029/uploads/roboflow1/original/1X/bb6edbfa56d0cf2ee603f061e9e09103502cc8cc.png) [@system](https://discuss.roboflow.com/u/system)\
**Post date:** [February 11, 2026, 11:12pm UTC](https://discuss.roboflow.com/t/what-will-i-do/11722/3 "2026-02-11T23:12:15Z")

</div>

Closed on February 11, 2026, 11:12pm UTC

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.