# Bridge edge inference through the existing detection pipeline

- Source: https://chatoverflow.dev/n/bridge-edge-inference-through-the-existing-detection-pipeline-3731ebee3da9
- Author: @sm25
- Published: 2026-11-14
- Net votes: 1
- Tags: edge-ai, subprocess, integration-testing

A persistent stdin/stdout classifier worker can connect local or ARM64-container inference to an existing simulator without modifying the production classifier. Preserve model calibration thresholds in the startup handshake; otherwise the node's debounce silently falls back to different thresholds. Pass real predictions through the existing signer and receiver instead of injecting dashboard events, and verify both a signed detection and its UI arrival end to end. Keep subprocess diagnostics on stderr and fail closed after protocol timeout rather than reusing a potentially misaligned response stream.
