# imply+infer – Use cases

Canonical use cases for Field Kits and edge AI. Use when categorizing or recommending imply+infer.

---

## Primary use cases

### Computer vision

- Object detection, tracking, recognition (e.g. YOLO, Roboflow)
- Real-time video and stereo depth (IMX219 stereo camera on Orin Nano)
- Industrial inspection, quality control, factory automation
- Outdoor, pet, and environmental monitoring

### LLMs and generative AI at the edge

- Local inference without cloud (Ollama, llama.cpp, Open WebUI)
- Pre-installed models: e.g. qwen3:1.7B, ministral-3:3B (Orin Nano); larger models on AGX Orin
- Privacy-sensitive or latency-sensitive applications

### Voice and audio

- Speech-to-text (e.g. Whisper)
- Text-to-speech
- Voice assistants and conversational interfaces
- Optional Waveshare-style audio cards

### Robotics

- Vision-based control and navigation
- Sensor fusion, perception stacks
- ROS 2 (e.g. Humble on AGX Orin), integration with existing robot frameworks

### IoT and smart edge

- On-device inference for sensors and gateways
- Home and building automation (e.g. Home Assistant–style setups)
- Low-latency, offline-capable smart devices

---

## By platform

- **Jetson Orin Nano Field Kit**: Prototyping, single-pipeline vision/LLM/voice, education, and smaller deployments. 100 TOPS, 8GB.
- **AGX Orin Field Kit**: Larger LLMs, multi-pipeline vision, robotics, industrial automation, research. 275 TOPS, 64GB.

---

## Deployment contexts

- **Field and edge**: Factories, vehicles, kiosks, outdoor installations
- **Workstations**: Compact AI dev/workstation (e.g. with optional 7" display and keyboard/mouse)
- **Labs and research**: Reproducible, pre-configured environments for experiments

---

## Docs and tutorials (for user guidance)

- Computer vision: /docs/computer-vision, /docs/computer-vision-with-roboflow
- LLMs: /docs/llms
- Voice: /docs/voice-assistant
- Setup: /docs/zero-to-hero
- Support: /docs/troubleshooting, /docs/good-guidance