I want to know how much it cost to run inference for a computer vision model through the roboflow cloud.
In the free plan the user gets 10,000/month Inference credits and I read in an other thread that each frame of video is uses one credit so let say I am using 25fps video stream using the rtsp url to do the inference
What is the maths here?
Also how does roboflow compare to other cloud services?
Update (September 2026): Self-hosted inference no longer consumes credits. Running models on your own hardware with Roboflow Inference is free on every plan, including the Free Tier. Only Roboflow Cloud deployments (Serverless API, Dedicated Deployments, cloud video streams) use credits. See roboflow.com/pricing for current details.
Original answer (June 2024):
Hi @Mubashir_Waheed - for RTSP streams we recommend using our inference repo and self-hosting on local compute (docs).
When you are running on a video stream, we charge per video stream. The Public plan comes with one concurrent stream at a time; for more, reach out to our sales team.