Use a webcam with an MCP assistant

Check your browser and camera

Use a browser with camera permission support on an HTTPS page. Before connecting an assistant, check that the webcam appears in the browser's camera selector or can provide a preview in a camera page. A USB camera must expose a standard webcam feed; a camera that works only in its maker's separate application may not be selectable by the browser.

Your AI client needs remote MCP, Streamable HTTP, OAuth, and native image results. Add https://letitsee.com/mcp, complete sign-in and consent, and note which account you used. You will sign in to the capture page with that same account. Camera and sign-in permissions are handled by your browser and account; the assistant cannot grant either one.

Choose and frame the webcam

When ready, open the camera page and sign in with the matching account. On desktop, choose Use camera, approve the browser permission prompt, and select the intended webcam from the camera selector. The browser can list several sources, such as an internal camera, external webcam, or virtual device. Select and verify the one you want in the preview; do not assume the default is correct.

Adjust the camera's physical position, focus, or lighting yourself. Place the subject in the middle of the frame, reduce glare, and keep the detail large enough to inspect. Leave the capture page open. Once you have framed the view, ask the assistant for one fresh photo and describe a specific visual question.

Troubleshooting a missing or wrong feed

If the browser selector is empty, confirm that the webcam is connected, recognized by the operating system, and not exclusively held by another application. Recheck the browser's camera permission for the page. If a feed appears but shows the wrong camera, choose the intended source in the page rather than relying on camera order. A phone or another browser page signed into the same account can replace the active page, so reconnect only when you intend to use this one.

Example prompt and limitations

A capture is one still image per request. LetItSee does not continuously send the webcam feed, record microphone or system audio, or provide a photo archive. Stop or close the camera page when finished. The assistant cannot start the camera, change the selected device, or choose what is in frame. Check the privacy policy for image handling details, and do not use visual interpretation for emergency or safety-critical decisions.

Set up without starting a capture

Set up your MCP client with the full tool reference. The separate capture page can replace the active capture page for your account. It does not take a photo by itself. When your source is ready, try:

Take one photo of the model on my desk. Compare the left bracket with the reference I uploaded and tell me what is visibly different.