Chainsaws
Illegal logging
ELOC’s chainsaw detector is proven and running in the field today. The moment a saw starts up within earshot, a patrol can be alerted — long before a tree falls.
In full deployment
Open-source · solar-powered · listening 24/7
AI ears for the rainforest
ELOC is a solar-powered bioacoustic recorder that runs AI directly on the device to hear chainsaws, gunshots and elephant rumbles — then alerts rangers in seconds over long-range radio, with no cellular network needed.
Built by the International Elephant Project to protect Sumatra's forests — and adaptable to any sound, anywhere.
Detection · live
eloc-00246
Chainsaw detected
0.94
Sent over LoRa → TTN ✓ 1.2s ago
130+
Devices deployed
Sumatra · Malaysia · Sri Lanka
Live
Chainsaw detection
In full field deployment
3
Countries
Sumatra · Malaysia · Sri Lanka
0
Cellular needed
Long-range LoRa radio
Why it matters
The forest is vast. The threats are silent.
The forest is vast. The threats are silent.
Until now.
The Sumatran elephant is critically endangered — its numbers falling, hemmed in by shrinking forest. Poaching, illegal logging and human–elephant conflict claim ground every day, across terrain no patrol can watch around the clock.
Traditional monitoring is expensive, slow, and impossible to scale across remote rainforest. ELOC changes the maths: low-cost, solar-powered sensors that listen continuously and raise the alarm the moment something is wrong — turning a silent forest into an early-warning network.
Illegal logging
Chainsaws strip protected forest. A single detection can put a patrol on a logger before the canopy comes down.
Poaching
Gunshots echo through the canopy. Real-time alerts help rangers reach the scene while it still matters.
Human–elephant conflict
Elephants raiding crops drives retaliation. Hearing them approach gives communities time to respond safely.
How it works
From a sound in the forest to a ranger's screen
Four steps, fully autonomous, running on a device the size of your hand — out in the jungle for weeks at a time.
-
Listen
A solar-powered device records the forest 24/7 in high-fidelity 24-bit audio, day and night, rain or shine.
-
Detect
An on-device AI model (Edge Impulse) recognises chainsaws, gunshots and elephant rumbles in real time — no internet required.
-
Alert
Detections fly out over long-range LoRa radio to The Things Network in seconds — reaching rangers where there is no cell coverage.
-
Act
Alerts surface on a live web dashboard — maps, device health and a detection timeline — so teams can respond fast.
What it hears
Three threats. One device. Real-time.
ELOC runs a sound-classification model on the device itself — so it reacts in the forest, not in the cloud.
Gunshots
Poaching
A gunshot is a sharp acoustic signature ELOC is trained to recognise. Real-time alerts help rangers reach poaching activity while there is still time to act.
Detection class
Elephant rumbles
Conservation & HEC
Elephants communicate in low-frequency rumbles. Hearing a herd approach supports conservation research and gives communities warning to avoid conflict.
Detection class
…and whatever the forest needs next
ELOC is a general bioacoustic platform, not a single-purpose alarm. The AI model can be retrained on new sounds — other species, vehicles, voices — so the same hardware adapts to new conservation challenges anywhere in the world.
Powered by
ESP32 · 240 MHz
The technology
A research-grade listening post, built to be cheap
ELOC packs edge AI, long-range radio and weeks of solar-backed autonomy into a low-cost, open-source device — engineered for the realities of the rainforest.
On-device edge AI
An Edge Impulse / TensorFlow Lite Micro model runs on the ESP32 itself — classifying sound in real time with no connection to the cloud.
Solar + smart sleep
Runs on a small solar panel and rechargeable battery. Duty-cycle deep sleep stretches battery life roughly 10–15× for weeks of unattended operation.
Long-range LoRa
An SX1262 LoRa radio relays detections over kilometres to The Things Network — independent of any cellular tower or mobile signal.
Studio-grade audio
A 24-bit MEMS microphone (ICS-43434) captures the full forest soundscape — high-fidelity recordings for detection and later analysis.
Tamper aware
An onboard accelerometer flags tampering or movement, and the device can be woken with a tap — useful when someone interferes with a unit.
Field-ready storage & GPS
High-speed microSD keeps every recording; optional GPS provides location and auto time-zone — all configurable on-site over Bluetooth.
One connected system
Device → radio → dashboard
Three open-source components work together to get a detection from deep in the forest onto a ranger’s screen.
01 · Firmware The device
The solar-powered recorder in the field. C/C++ on ESP-IDF + Arduino: captures audio, runs the AI detector, sends alerts over LoRa, and serves a Bluetooth config interface.
- On-device AI inference
- LoRaWAN alerts
- Records to microSD
02 · Android Control Panel app
A field tech’s pocket control panel. Connects to a device over Bluetooth to configure recording, detection thresholds and radio — and to read live status. No laptop, no signal needed.
- Bluetooth setup on-site
- Start/stop & status
- Works offline
03 · For rangers & admins Web dashboard
The operational picture. A live map of every device, alert timeline, battery and GPS health — fed in real time from The Things Network and the app.
- Live device map
- Alert history & replay
- Role-based access
Under the hood Full technical specifications
ELOC 3.0 hardware & software. Tap to expand.
Full technical specifications
ELOC 3.0 hardware & software. Tap to expand.
Processing
- Processor
- ESP32-WROVER — dual-core, up to 240 MHz, with PSRAM
- Flash storage
- 16 MB on-board + high-speed microSD (SDIO)
- On-device AI
- Edge Impulse / TensorFlow Lite Micro
Audio
- Microphone
- TDK InvenSense ICS-43434 I²S MEMS (24-bit)
- Sample rate
- Configurable ~4–51.6 kHz (16 kHz typical)
- Alt. mics supported
- PUI DMM-4026-B-I2S-R · SPH0645
Connectivity
- Long-range radio
- SX1262 LoRa · LoRaWAN 1.1 (OTAA, The Things Network)
- Default region
- AS923 (Sumatra / SE Asia)
- Local config
- Bluetooth Classic (SPP), JSON commands
- Positioning
- Optional GPS (ATGM336H) — time sync & auto time-zone
Power & sensors
- Power
- Solar panel + rechargeable LiPo / LiFePO₄, USB charging
- Battery life
- 10–15× extension via duty-cycle deep sleep; unlimited with solar
- Sensors
- LIS3DH accelerometer (tamper/tap), battery-voltage monitor
Software
- Firmware
- C/C++ — ESP-IDF + Arduino, FreeRTOS, PlatformIO
- Updates
- Over-the-air firmware update via SD card (dual-OTA partition)
- Licence
- Open source · MIT
In the field
Already listening across three countries
ELOC devices are deployed with conservation teams on the front line — and because the platform is open and low-cost, a single forest can be blanketed with sensors rather than dotted with a few.
Sumatra, Indonesia
Leuser Ecosystem · with FKL ranger teams
Malaysia
Forest monitoring deployments
Sri Lanka
Human–elephant conflict zones
What it's used for
- Anti-poaching patrols
- Illegal-logging detection
- Human–elephant conflict mitigation
- Behaviour & population monitoring
- Acoustic habitat research
- AI training-data collection
130+
devices deployed
3
countries
24/7
listening
Open source
Built in the open, so anyone can protect their forest
Open-source hardware, firmware, app and dashboard — all under the MIT licence. Inspect it, build it, deploy it, improve it. Conservation technology shouldn't be a black box.
Repository: LIFsCode/ELOC-3.0 · Licence: MIT
- Open hardwareESP32 design & BOM you can build yourself
- Open firmwareC/C++ on ESP-IDF + Arduino, on GitHub
- Open app & dashboardAndroid + React, free to fork
- Community-drivenBacked by conservation engineers & researchers
A conservation collaboration
Developed by the International Elephant Project and the Sumatran Elephant Conservation Initiative, with engineer Stephan Stange and field teams from Forum Konservasi Leuser and a network of academic collaborators.
Get involved
Help the forest be heard
Whether you build, deploy, research or fund — there's a way to be part of ELOC.
Build & deploy
Grab the open-source hardware, firmware and app from GitHub and put a device in the field.
Open the repoConfigure a device
Download the ELOC Control Panel app to set up and monitor devices over Bluetooth.
Get the appPartner or support
Researchers, conservation teams and funders — get in touch with the International Elephant Project.
Contact us