If I try to make an inference with a local RTSP, I still need to run API_KEY and server, right?

[YoungjaeDev]

Search before asking

• [X] I have searched the Inference issues and found no similar feature requests.

Question

Due to the roboflow inference structure, even if I simply use a local RTSP_STREAM, do I still need to start the server and enter the API_KEY? Or is there another way?

Additional

No response

[grzegorz-roboflow]

Hi @YoungjaeDev, you do not need to provide API_KEY if you are using foundational models. Consider below example:

source /path/to/venv/where/inference/is/installed/bin/activate
inference server start
inference infer --input /path/to/file.jpg --model_id yolov8m-seg-640
# yields {'time': 0.24999908400002369, 'image': {'width': 3000, 'height': 4000}, 'predictions': [ ... ]}

However as soon as you want to use a non-foundational model (i.e. model you trained) you need to provide API key:

source /path/to/venv/where/inference/is/installed/bin/activate
inference server start
inference infer --input /path/to/file.jpg --model_id my_private_model --api
-key <secret>

[grzegorz-roboflow]

Hi @YoungjaeDev, I will go ahead and close this issue. Please reopen if you would like to share more context.