Integrating Blue Iris with CodeProject.AI Server introduces a two-tier system:
Adjust the confidence threshold (e.g., 70%). If the AI is only 50% sure it’s a person, it might be a shadow. Increasing this threshold reduces false alerts. Conclusion codeproject blue iris verified
: Force Blue Iris to send the lower-resolution substream (e.g., 1080p or 720p) to CodeProject.AI for analysis, rather than the heavy 4K main stream. Accuracy remains high, but processing times drop significantly. Integrating Blue Iris with CodeProject
CPU usage spikes to 100%; inference time is > 500ms. Fix: In CodeProject.AI Server dashboard ( http://localhost:32168 ), check System Info . If your NVIDIA GPU is not listed, install the correct CUDA toolkit (v12.x). Restart the AI server. Conclusion : Force Blue Iris to send the
Integrating into your Blue Iris surveillance setup has become the gold standard for home security enthusiasts. Moving away from legacy systems like DeepStack, this combination offers "verified" event detection, which uses locally hosted artificial intelligence to confirm exactly what is happening in your camera's frame before sending an alert. Why "Verified" Matters
By combining Blue Iris NVR software with the open-source CodeProject.AI Server, users can transform basic motion triggers into verified alerts for people, vehicles, and custom objects.