Sample price, not a volume price.
Request a sampleYou already own the screens and the playback system. We build the digital human, optimise it to render locally on our hardware, and connect it to the signage system you run today.
Open this page on a phone or a computer, tap the microphone and speak to the avatar directly — the same system that runs on the signage screens.
We build your 3D or 2D digital human, render the model locally on our hardware, and connect the back end to your own content and advertising system. A shopper walks up, it says hello. A shopper asks, it answers.
What the avatar says and plays is decided by the systems you already run. Our job is to plug into them, not to make you switch.
No. Any screen with an HDMI port can be upgraded. The hardware plugs straight into HDMI and draws Type-C power from the screen's own USB port, and your existing playback device stays where it is. Sites can be retrofitted one screen at a time.
A subscription bills monthly and by the minute, keeps the character and model on somebody else's platform, and the invoice grows with every interaction. Stop paying and the avatar is gone, and because frames come from the cloud it stops the moment the network drops. This is made once and used forever. The model is handed over and installed on your own device, interactions cost nothing extra, and local rendering keeps drawing frames on a weak connection. The cost is a one-off avatar build, hardware priced per site, and optional integration and knowledge-base setup.
Our own AI TV stick, on Android 14 with 1080P output and 4K source decoding, Wi-Fi 6 dual-band and Bluetooth 5.4. The model is optimised to the hardware and rendered locally in real time rather than streamed from a cloud. Microphone array, camera or presence sensing are specified per site.
Yes. A brand character, a likeness of a real person or an original IP can all be built, in either 3D or 2D. A 3D avatar is fully dimensional, turns and appears full body, with modelling, rigging, expressions, lip sync, stance and gestures all produced. A 2D avatar is faster and cheaper, appears from the waist up, and can be illustrated or photoreal. Appearance and clothing, voice timbre and pace, personality and script, multilingual voicing and the background scene are all customisable.
A 2D model takes roughly 2 weeks, a 3D model 4 to 6 weeks depending on complexity. Hardware mass production runs 30 days. We suggest proving one site first, then replicating hardware by screen count.
Yes. Creative and scheduling stay in your playback backend. Product, price, promotion and store information are imported and become answerable. Membership, stock, queueing and process systems connect over APIs. Interaction counts, frequent questions and captured leads flow back into your backend.
With nobody present it returns to the advertising loop. When it senses somebody approaching it switches from standby to a greeting. One screen, two modes.
Hardware mass production MOQ is 1,000 pcs, with pilot quantities and samples quoted separately. CE, FCC, UL and UKCA, with a 2-year whole-unit warranty.
Pricing depends on your requirements, such as 2D or 3D, appearance, scripts, knowledge base and integration scope. Tell us the use case and we quote per project. You can try the live demo on this page first.
Yes. The digital human runs on the HDMI AI TV stick we supply, or on our projectors, and an external USB directional microphone adds voice interaction. Requirements differ by site, so the setup and price are quoted per project.
Tell us how many screens you have, what kind of sites they are and the character style you want. We will come back with an avatar proposal and full pricing — or put one stick on a single screen for a trial run.
PINGOOD TECHNOLOGY CO. LIMITED (trading as AI HW Stack) is a Shenzhen-based AI hardware ODM/OEM. We deliver the complete stack — industrial design and tooling, mainboard, Android firmware, the ASR + LLM + TTS pipeline, the companion app and the cloud backend — as one order.
Technology we build on: Android TV, Azure AI Speech, OpenAI speech-to-text, Alibaba Model Studio, Cloudflare Workers.
Published 2026-09-01 · · Contact: kristy@aihwstack.com
