
SMT / Electronic Assembly
Printing machine / SPI / reflow soldering / AOI screen-parameter acquisition and correlation analysis.
View scenario
Video OCR data acquisition and analog keyboard & mouse reverse control for legacy and closed devices, bringing data islands into digitalization.
When a device has no open interface and cannot be retrofitted, the AIS-T100 turns key parameters into structured data by “reading the screen”.
The AIS-T100 adopts an Intel high-performance computing chip and embeds a range of video and image algorithms to enable local data processing, while supporting analog keyboard & mouse reverse control. It connects directly to the legacy host’s VGA signal cable, captures the video signal in real time and outputs it losslessly to the original display via VGA OUT, keeping the original host screen uninterrupted.
The original host VGA signal is converted to a digital signal by the module, enabling second-level frame extraction to images. By mouse frame-selection of the parameters to be collected, the image coordinate position is auto-recognized; a preset image-binarization algorithm allows threshold adjustment to optimize character-recognition accuracy (up to 99%), completing frame-selection, recognition and annotation of a single parameter within ten minutes.
On top of acquisition, the AIS-T100 supports multi-page adaptation, model management, edge-computing cleansing and alarming, and reports upward to various cloud platforms and systems via protocols such as MQTT / TCP / IP / Modbus — bringing even “invisible” legacy equipment online for digitalization, and serving as the key node for networking island-type data systems.
Core capabilities built around “capture accurately, control reliably, connect upstream”.
VGA IN capture + VGA OUT lossless loop-out keeps the original display uninterrupted, making deployment invisible to production.
Up to 99% recognition accuracy; mouse frame-selection of parameters, modeling completed within ten minutes, with a preset binarization algorithm and adjustable threshold.
Adaptively matches different pages by image characteristics and loads the corresponding parameter settings, adapting to automatic multi-page switching.
Build, import and export engineering parameter models by target device model, enabling reuse across similar-device projects.
Simulates keyboard and mouse via USB and PS/2 interfaces, supporting software scripts to operate the original host in reverse.
Local edge cleansing, alarming and triggering; supports MQTT / TCP / IP / Modbus northbound reporting.
Complete hardware and software parameters from the product specification sheet.
| CPU | Intel Elkhart Lake J6413 Quad Core Processor, 1.8GHz, TDP 6W |
|---|---|
| GPU | Intel CPU Integrated Gen.11 Graphics Engines |
| Memory | 8GB DDR4-3200 standard, up to 32GB supported |
| Storage | 256GB SSD standard, up to 2TB supported |
| Network | 2 × independent 2.5Gb RJ45; WiFi 2.4G; 4G/5G optional (miniPCIe) |
| Serial Port | 4 independent serial ports, supporting RS232 / 422 / 485 (BIOS setting) |
| USB | 4 × Type-A: 3 × USB3.1 + 1 × USB2.0 |
| Display | HDMI + DP dual display, supports 4096×2160@60Hz |
| Video Capture | VGA IN capture + VGA OUT loop-out, second-level frame extraction to images |
| OCR Accuracy | Up to 99%, frame-selection, recognition and annotation within ten minutes |
| Underlying OS | Linux Ubuntu 22.04 |
| Power | DC +10V ~ 30V input, typical consumption 13W |
| Operating Environment | -20 ~ +60°C, humidity ≤95% |
| Protection Rating | IP65, aluminum-alloy industrial-reinforced material |
| Installation | Wall-mount fixed, DIN-rail mounted |
| Dimensions | 150 × 107 × 39 mm (L×W×H) |
Focused on solving the legacy-equipment networking challenge of “no interface, no retrofit, scattered data”.

Printing machine / SPI / reflow soldering / AOI screen-parameter acquisition and correlation analysis.
View scenario
Closed system with no external interface: T100 auto-cycles screens to feed isolated data into the IOT platform.
View case
Monitoring and batch tracing of key process parameters of old production-line equipment.
View scenarioHow the AIS-T100 differs from competitor acquisition approaches that “must retrofit the device / must open the protocol”.
No need to open the device interface or retrofit it; simply “read the screen” to collect, fitting a large number of legacy devices.
VGA loop-out keeps the original display working normally, making deployment invisible to production and low-risk.
Up to 99% recognition, ten-minute frame-selection modeling, adjustable binarization threshold for complex screens.
Local cleansing, alarming and triggering, reducing cloud bandwidth and server load.
Analog keyboard & mouse operate the original host in reverse, closing the acquisition–control loop.
IP65, wide temperature, aluminum-alloy housing, suited to harsh industrial sites.
Real-world results from AIS-T100 at customer sites; click to view the full case.
Clarifying the division of labor with APM-G220 / AIS-G160.
✅ Applicable to: screen-parameter collection and analog keyboard & mouse reverse control for legacy industrial equipment and closed island systems with no open interface.
❌ Not applicable: desktop-level remote operation & maintenance of PC-controlled equipment (choose APM-G220); protocol-layer deep acquisition of dedicated controllers (choose AIS-G160).
Combination recommendation: AIS-T100 can be combined with AIS-G160 — the former solves “screen-layer” data acquisition and the latter solves “protocol-layer” data acquisition, together forming a plant-wide data aggregation solution.
High-frequency questions about legacy-device screen capture.
Use the AIS-T100 video OCR gateway: connect to the injection machine controller’s VGA / HDMI signal and OCR-recognize the on-screen parameters — no interface needed to complete data acquisition and networking.
Yes — this is exactly the core scenario of the T100. As long as the device has a displayable screen interface, the T100 turns on-screen parameters into structured data via video capture + OCR recognition, with zero retrofit and no production interruption.
The T100 uses deep-learning OCR. For industrial LED / LCD digit tubes, dot-matrix and character displays, combined recognition accuracy reaches up to 99%, and parameter frame-selection modeling can be completed within ten minutes.
The T100 converts on-screen parameters into structured data at the edge and uploads via standard protocols such as MQTT / Modbus TCP, connecting to mainstream MES / IOT platforms and feeding a unified data model.
AIXOT’s three gateways each play their role; combined, they cover plant-wide equipment digitalization.