# jomjol/AI-on-the-edge-device

Easy to use device for connecting "old" measuring units (water, power, gas, ...) to the digital world

Repository: https://github.com/jomjol/AI-on-the-edge-device
Canonical: https://ross.abutalabs.com/products/ai-on-the-edge-device
Homepage: https://jomjol.github.io/AI-on-the-edge-device-docs/
Language: C++
License: NOASSERTION
License Family: other
Last push: 2026-07-03T21:50:32+00:00

## Health v2 (maintenance only)
Score: 72/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 90, release rhythm 33, longevity 100
- inputs: {"age_days": 2227, "days_push": 61, "days_rel": 234, "gap_med": 302, "n_releases_24m": 2}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 8624, forks 899 (observed 2026-08-28T04:10:23.891857+00:00)

## What it is
A firmware application for ESP32-CAM boards that uses TensorFlow Lite CNNs on-device to digitize analog utility meters (water, gas, electricity). It captures camera images, aligns and extracts ROIs, recognizes digits, and exposes results via MQTT, REST API, InfluxDB, and Home Assistant.

## Use cases
- read analog water meter readings automatically
- digitize old electricity meter with ESP32 camera
- send gas meter values to Home Assistant
- on-device OCR of utility meter digits
- monitor water consumption with MQTT
- convert non-smart meters to smart meters

## When to choose
- you have an analog meter and want cheap, fully local digitization without cloud
- you want Home Assistant/MQTT/InfluxDB integration out of the box
- you prefer edge inference on a sub-10-EUR device

## When to avoid
- your meter is already digital/smart with a data interface
- you need high-accuracy recognition on meters with unusual dials not covered by trained models
- you cannot physically mount a camera near the meter

## Facets
- artifact type: application
- maturity: active
- function: machine-learning, computer-vision, image-processing, iot, http-server, monitoring
- domain: iot, computer-vision, machine-learning, self-hosted
- platform: embedded, iot, self-hosted
- tags: esp32-cam, meter-reading, tensorflow-lite, edge-ai, home-assistant, mqtt, analog-meters, water-meter, power-meter, gas-meter

## Member repositories
- jomjol/AI-on-the-edge-device (main) score 72

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:23.891857+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:25:52.422043+00:00, confidence not recorded.
  - readme: https://github.com/jomjol/AI-on-the-edge-device (fetched 2026-08-28T04:10:23.891857+00:00, sha 5c66fa015dbf)
  - homepage: https://jomjol.github.io/AI-on-the-edge-device-docs/ (fetched 2026-08-29T08:26:09.848536+00:00, sha d5cc63b8574d)
- Data as of 2026-08-30T08:39:29.467469+00:00.