- Description
- Specifications
- Download
- FAQ
- Reviews
Overview
InHand Networks Mo 68A AI Single Board Computer
The Mo 68A is an edge-AI single board computer built on the TI AM68A vision processor. With dual C7x DSPs and a deep-learning accelerator delivering 8 TOPS of on-device inference, an on-chip ISP with VPAC / DMPAC vision acceleration engineered for real-world camera scenes, dual 4-lane MIPI CSI-2 camera inputs and 4K60 video codec, and a standard SBC form factor compatible with the HAT ecosystem, the Mo 68A turns multi-camera AI vision ideas into deployable edge devices — at a fraction of the cost and power of traditional AI hardware. Running Debian Linux with an open SDK and TI TIDL support for TFLite/ONNX models, it is built for developers who want to go from prototype to production — the open-source way.
Features
8 TOPS of Edge AI — Inference Where the Camera Is
Two C7x DSPs with a deep-learning accelerator deliver 8 TOPS of deep-learning performance on-device. Detection, classification and segmentation models run locally — no cloud round-trips, no bandwidth bills, no privacy exposure — with headroom for multiple camera streams and heavier models. The TI TIDL toolchain deploys your existing TFLite and ONNX models with minimal effort.
A Vision Processor, Not Just an SBC
Unlike general-purpose boards, the AM68A integrates an on-chip ISP with VPAC vision acceleration — hardware WDR, lens distortion correction and RGB-IR support — plus DMPAC depth acceleration and 4K@60fps video encode/decode. Backlit entrances, high-contrast sites and low-light scenes are handled in silicon before your model ever sees the frame, lifting real-world detection accuracy.
Raspberry Pi Form Factor, Industrial AI Core
Standard 85 × 56 mm SBC footprint with a 40-pin HAT-compatible header (GPIO / I²C / I²S / SPI / UART / PCM). Mounts in mainstream SBC enclosures and works with the vast ecosystem of HAT expansion boards — PoE HATs, relays, sensors, displays — so prototyping hardware is never a blocker.
Built for Developers — Open SDK, Open Source
Ready-to-flash Debian system images, an open SDK for custom builds, and TIDL model deployment examples — flash and run your first inference in minutes. Developer documentation and downloads are available in the InHand Resource Center, and the Mo 68A GitHub repository is on its way. No proprietary runtime in the way: develop in Python or C/C++ with OpenCV, GStreamer and NumPy, manage packages with apt, and access the board over SSH. Your code, your system, your product.
Connectivity for Real Deployments
Gigabit Ethernet for reliable backhaul, four USB 3.0 Type-A ports for peripherals and additional cameras, mini DP plus MIPI DSI for local display, a PCIe 3.0 interface for high-speed expansion, and dual 4-lane MIPI CSI-2 for direct multi-camera input — everything a multi-camera edge vision terminal needs, nothing it doesn't.
Secure by Design
Secure boot, Arm TrustZone, OP-TEE trusted execution and a hardware AES-256 crypto accelerator protect firmware, models and data end to end — essential when devices are deployed unattended in public spaces and your AI model is your IP.
From Prototype to Production
USB-C 5V/5A power, 25 W maximum consumption and a 4-pin PWM fan header keep deployment simple under sustained loads. Two memory options (4GB / 8GB) let you match cost to workload — validate on a single board today, order production volumes tomorrow, on the same platform and the same software.
Applications
From a developer's desk to a thousand deployed sites — one compact platform for vision AI at the edge.
Developer Resources — Open Source on GitHub
We develop in the open. The Mo 68A developer environment — system images, SDK, AI examples and documentation — is being prepared for release on GitHub, and downloads are already available in the InHand Resource Center.
System images · EdgeAI SDK · TIDL model examples (TFLite / ONNX) · GPIO & device-tree docs · Issue tracker — the repository is being prepared. Meanwhile, firmware and documentation are available in the InHand Resource Center (see the Download tab).
Why Mo 68A — Compare Before You Build
General-purpose SBCs leave AI to add-ons; high-end AI kits cost several times more. The Mo 68A puts a dedicated multi-camera vision AI pipeline in a standard SBC form factor — right-sized for dual-camera, high-resolution edge inference.
| Specification | Raspberry Pi 5 | Jetson Orin Nano |
Typical RK3588 SBC | InHand Mo 62A |
InHand Mo 68A |
|---|---|---|---|---|---|
| Positioning | General-purpose SBC | High-perf AI dev kit | Multimedia SBC | Edge AI vision SBC | Multi-camera vision AI SBC |
| AI acceleration | — (requires AI HAT) | Up to 40 TOPS | 6 TOPS NPU | 2 TOPS (C7x DSP + MMA) | 8 TOPS (2× C7x DSP) |
| AI toolchain | Via HAT vendor | TensorRT | RKNN | TIDL · TFLite / ONNX | TIDL · TFLite / ONNX |
| CPU | 4× Cortex-A76 @2.4GHz | 6× Cortex-A78AE | 8× (A76+A55) | 4× Cortex-A53 @1.4GHz | 2× Cortex-A72 @2.0GHz |
| Camera input | 2× CSI (mini) | Via carrier board | MIPI CSI | 1× 4-lane MIPI CSI-2 | 2× 4-lane MIPI CSI-2 |
| Vision ISP | Basic | Yes | Yes | ISP + VPAC (WDR · LDC · RGB-IR) | ISP + VPAC + DMPAC · 4K60 codec |
| Video | 4K60 decode | 4K60 | 8K decode | micro HDMI out | 4K60 encode/decode · mini DP |
| Expansion | 40-pin GPIO | 40-pin GPIO | 40-pin GPIO | 40-pin HAT-compatible | 40-pin HAT-compatible · PCIe 3.0 |
| OS | Raspberry Pi OS | JetPack (Ubuntu) | Linux / Android | Debian 13 · open SDK | Debian 13 · open SDK |
| Power input | USB-C 5V/5A | 7–20V DC | 12V DC | USB-C 5V/5A · 25W max | USB-C 5V/5A · 25W max |
* Competitor specifications are collected from public sources for reference only. Please refer to each vendor's official documentation for final data. Need a lower-cost single-camera option? Check out the Mo 62A.
In the Box
Why InHand
InHand Networks has built industrial connectivity and edge computing devices for over two decades, deployed in energy, transportation, manufacturing and smart-city projects worldwide. The Mo series brings that industrial engineering discipline to edge AI: documented hardware, maintained software, and a supply chain you can plan a product on.
Hardware
| Processor | TI AM68A, 2 × Arm Cortex-A72 @ 2.0 GHz |
| AI Accelerator | 2 × C7x DSP + Deep Learning Accelerator, 8 TOPS |
| ISP / Vision | On-chip ISP + VPAC (RGB-IR, WDR, LDC) + DMPAC |
| RAM | LPDDR4 8GB (default) / 4GB |
| Storage | Micro SD card |
| Ethernet | 1 × Gigabit Ethernet |
| USB | 4 × USB 3.0 Type-A |
| PCIe | 1 × PCIe 3.0 |
| Display | 1 × mini DP; up to 2 × 4-lane MIPI DSI |
| Camera | Up to 2 × 4-lane MIPI CSI-2 |
| Audio | I²S (via 40-pin connector) |
| Expansion | 40-pin header: GPIO / I²C / I²S / SPI / UART / PCM, 3.3 V logic, HAT-compatible |
| Debug | 1 × TTL UART console |
| Fan | 1 × 4-pin fan header (5V / PWM / GND / TACH) |
| Button | 1 × Reset |
| LED | PWR, STATUS |
| RTC | Supported, with battery backup |
Power & Mechanical
| Power Input | USB Type-C, 5V / 5A DC |
| Power Consumption | 25 W (max) |
| Dimensions | 85 × 56 mm |
| Weight | 53 g |
| Operating Temperature | 0 ~ 50 °C |
| Storage Temperature | -20 ~ 70 °C |
Software
| Operating System | Debian 13 (Trixie), Linux kernel 6.12 |
| AI Runtime | TI TIDL — deploys TFLite / ONNX models |
| Vision SDK | TI EdgeAI SDK; V4L2 camera framework; DRM/KMS display |
| Development | Python, C/C++; OpenCV, GStreamer, NumPy; apt package manager |
| Security | Secure boot, Arm TrustZone, OP-TEE, hardware AES-256, watchdog |
| Open SDK | Custom system builds supported |
| Management | SSH remote access; SD card image flashing; UART console |
Models
| Model | RAM |
|---|---|
| Mo-68A-4G | 4GB LPDDR4 |
| Mo-68A-8G | 8GB LPDDR4 |
The complete Mo 68A developer environment — system images, SDK source, TIDL model examples and docs — is being prepared for release on GitHub.
What does the Mo 68A offer?
The Mo 68A is an 8-TOPS AI single-board computer for on-device, multi-camera vision inference at the edge. Key specs:
| Compute | 2× Cortex-A72 @ 2.0 GHz + 2× C7x DSP with deep-learning accelerator, 8 TOPS |
| Vision pipeline | On-chip ISP + VPAC (RGB-IR, WDR, LDC) + DMPAC; up to 2× 4-lane MIPI CSI-2; 4K60 video codec |
| RAM | LPDDR4 — 4 GB / 8 GB SKUs (Mo-68A-4G / -8G) |
| OS | Debian Linux |
| Connectivity | 1× Gigabit Ethernet, 1× PCIe 3.0 |
| Other I/O | 4× USB 3.0 Type-A, mini DP, up to 2× MIPI DSI, 40-pin HAT-compatible header, micro SD |
| Security | Secure Boot, Arm TrustZone, OP-TEE, hardware AES-256 |
| Power / size | USB Type-C 5 V / 5 A (25 W max), 85 × 56 mm, 53 g |
The 85 × 56 mm board and 40-pin layout are mechanically compatible with standard SBC enclosures and HAT accessories. Typical applications: AI vision boxes, intelligent cameras and other edge-AI inference terminals. The Mo 68A is a developer-oriented SBC managed via SSH and standard Debian tooling — it does not run InHand IEOS or DeviceLive.
Which AI frameworks and model formats are supported?
Models run through the TI Deep Learning (TIDL) runtime, which accepts TFLite (.tflite) and ONNX (.onnx) models. The on-chip ISP/VPAC handles RAW→RGB conversion, WDR and lens correction so the accelerator gets a clean input; pre- and post-processing in Python or C/C++ commonly use OpenCV and GStreamer.
What workloads is 8 TOPS enough for?
At 8 TOPS the Mo 68A is sized for multi-camera or higher-resolution workloads: multi-stream object detection and tracking, defect inspection, OCR, people/vehicle analytics and pose recognition — with 4K60 encode/decode for recording or streaming alongside inference. For lighter single-camera workloads at a lower cost, take a look at the 2-TOPS InHand Mo 62A.
How do I use the 40-pin GPIO header?
The 40-pin HAT-compatible header exposes GPIO, I²C, I²S, SPI, UART and PCM at 3.3 V logic. Standard SBC HAT accessories mount mechanically — verify electrical compatibility (3.3 V) before connecting. Alternate pin functions are configured via device-tree overlays; see the Hardware Interface Description in the Download tab for the full pinout.
How do I connect cameras, displays and high-bandwidth peripherals?
Two 4-lane MIPI CSI-2 connectors take direct camera input for single- or dual-camera setups. The mini DP output (plus up to two 4-lane MIPI DSI interfaces) drives local displays. Four USB 3.0 Type-A ports handle storage, adapters and additional cameras, and the PCIe 3.0 interface is available for high-speed expansion.
Does it work with Raspberry Pi HATs and enclosures?
Mechanically, yes — the 85 × 56 mm footprint and 40-pin layout match the standard SBC form factor, so mainstream enclosures and HATs mount directly. Electrically, the header is 3.3 V logic: check your HAT's voltage requirements before connecting.
No reviews yet — be the first to share your Mo 68A project experience.
