Developers — AI SENSE

// what.we_offer

Developer building blocks

⚡

Edge AI SDK

TinyML inference toolkit optimised for Raspberry Pi, NVIDIA Jetson, and ESP32 platforms. Deploy wildlife detection and crop health models with single-line inference calls — no cloud dependency required.

ONNX and TensorFlow Lite model export
Quantised models for 4x faster inference
Python 3.9+ and C++ bindings
Supports camera, microphone, and multi-sensor inputs

v0.4 Beta

🌐

REST APIs

Cloud endpoints for real-time inference, telemetry ingestion, and alert management. Integrate wildlife detection alerts, crop health scores, or farm sensor readings into any web or mobile application.

JSON over HTTPS — no SDK required
Endpoints: /detect, /health, /telemetry, /alerts
Rate limits: 500 req/day on free tier
Webhooks for real-time push alerts

Available

📦

Open Datasets

Annotated wildlife and agricultural datasets collected from field deployments in Maryland and Nepal. Suitable for model training, benchmarking, and academic research. Released under Creative Commons licences.

Wildlife Dataset v1.2 — 38,000 frames, 14 species
CropHealth Dataset v2 — multispectral, 12 stress indicators
Urban Farm Sensor Dataset — 6-month time-series
Bounding-box + polygon annotations included

CC-BY 4.0

// quickstart

Up and running in minutes

Install the Python SDK, point it at an image or live camera feed, and get structured detection results back in one call. No model configuration required — sensible defaults are baked in.

The same SDK works on a cloud server or a Raspberry Pi 4 with 4 GB RAM. Swap the model name for the task you need.

Python 3.9+

# Install: pip install aisense-sdk
from aisense import EdgeClient

# Initialise with your API key
client = EdgeClient(
    api_key="sk-your-key-here",
    model="wildlife-v2",
    device="cpu"  # or "cuda", "jetson"
)

# Run detection on a single image
result = client.detect(image_path="field.jpg")

# Structured output
for det in result.detections:
    print(det.label, det.confidence,
          det.bbox)

# Stream from live camera
with client.stream(source=0) as cam:
    for frame in cam:
        print(frame.detections)
          

// integration.steps

From zero to deployed

Get your API key

Contact us or fill in the access request form. We'll provision your key within one business day and share the SDK documentation.

Install the SDK

Run pip install aisense-sdk — supports Python 3.9+ and C++ via header-only bindings. Tested on Linux, macOS, and Raspberry Pi OS.

Deploy on edge

Flash the optimised model bundle to your edge device. Our ONNX export path supports Jetson Nano, Pi 4, and generic x86 servers. TensorFlow Lite targets ESP32 and Coral Edge TPU.

Stream to your dashboard

Push detection results to our cloud endpoint via MQTT or HTTP, or route directly to your own backend. Webhook support available for real-time alert delivery.

// tech.stack

Built on open standards

We build with widely-adopted tools so you never depend on proprietary lock-in.

🐍 Python 3.9+ ⚙️ C++ 17 📡 MQTT v5 🧠 ONNX Runtime 📦 TensorFlow Lite 🔭 OpenCV 4 🌐 REST / JSON 🪝 Webhooks 🐧 Linux / Raspberry Pi OS ⚡ NVIDIA Jetson 🪸 Coral Edge TPU ☁️ AWS / GCP / self-hosted

// get.started

Ready to build?

Request an API key, download a sample dataset, or reach out to our engineering team to discuss integration requirements.

Build with AI Sense