Far-field pickup with barge-in, on-device recognition, voice-driven equipment.
When kiosk speakers playback audio, raw microphones easily suffer severe acoustic feedback from their own output, blinding the device to user speech and forcing visitors to wait until lengthy prompts finish before speaking again.
Integrate hardware-grade AEC array boards that cancel the local playback reference directly at the chip level, capturing user barge-in utterances instantly even during 80dB loud speaker output to cut playback immediately.
Core Advantages
Public cloud LLM backends crash completely during internet outages or within isolated private networks, per-token API billing inflates long-term operating expenses, and enterprise sensitive inquiries risk data-compliance violations.
Deploy edge hosts with on-board NPUs running offline ASR, quantized Small Language Models (SLMs), and local TTS entirely on-premise, eliminating cloud reliance and locking round-trip response latencies under 1 second.
Core Advantages
Embodied robotic arms and service robots relying on touchscreen inputs fail to free operators' hands, while cloud-assisted voice control suffers variable network jitter that leads to delayed grasping motions and physical collisions.
Embed compact edge compute modules directly inside robots, parsing spoken commands into structured coordinate frames and trajectory codes on-device, dispatching motion controls over internal high-speed buses in tens of milliseconds.
Core Advantages
The configurator works through use case, voice tier, concurrency, on-device LLM, microphone setup and latency target to reach specific models.
Not sure which to choose? See the full selection guide →
Step by step, narrowing down to your recommended BOM
Wake/command words only; pair w/ your own host
| 硬件/物料名称 | 数量 | 型号/SKU | 说明 |
|---|---|---|---|
| XIAO ESP32-S3 Sense (Built-in Microphone) | 1 | 113991115 | - |
Single-lang local STT, no TTS. For meeting notes & voice memos.
| 硬件/物料名称 | 数量 | 型号/SKU | 说明 |
|---|---|---|---|
| ReSpeaker Lite (2-Mic Near-Field) | 1 | 107990273 | - |
| reRouter CM4 102032 | 1 | 110110205 | - |
Multi-lang ASR + TTS; cloud or local LLM optional.
| 硬件/物料名称 | 数量 | 型号/SKU | 说明 |
|---|---|---|---|
| ReSpeaker Lite (Near-Field ≤3 m) | 1 | 107990273 | - |
| reRouter CM4 102032 — Machine Voice Entry (Single-lang) | 1 | 110110205 | - |
Voice compute placement determines capability ceiling & per-unit BOM. Three common models:
Core Advantages
| Image | Hardware | Voice Concurrency (voice-only) | Voice Capabilities | Local LLM |
|---|---|---|---|---|
![]() | reRouter CM4 Series | 1 channel | Machine Voice | Cloud optional |
![]() | reComputer RK3576 Series | 1 channel | Simulated Voice | 4B–7B (w/ RK1828 card) |
![]() | reComputer RK3588 Series | 1 channel | Simulated Voice | 4B–7B (w/ RK1828 card) |
![]() | reComputer J30 Series | 1 channel | Natural Voice | ~2B |
![]() | reComputer J40 Series | 2–3 channels | Natural Voice | 4B–7B |
![]() | reComputer J50 Series | 3+ channels | Natural Voice | 7B–14B |
![]() | NVIDIA Jetson AGX Thor Developer Kit | 6+ concurrent (voice-only) | Realistic voice | 14B–32B |
The table selects the device for the Voice Host (ASR / TTS / Agent) block in the middle of the architecture diagram, listed by family: voice-only concurrency, voice capability, and local-LLM size per family. Exact models, configurations and accessories are picked in the configurator below (mic & speaker selection: next tab). Note: running a local LLM beside the voice pipeline reduces performance and adds latency; for high concurrency plus large models, choose a higher compute tier.
Core Advantages
| Image | Product | Type | Best for | Key specs |
|---|---|---|---|---|
![]() | ReSpeaker Lite | Pickup (near-field) | ≤ 3 m / desktop / single workstation | 2-mic array / on-board AI audio processing / USB · I²S |
![]() | ReSpeaker XVF3800 | Pickup (mid/far-field) | 3–5 m / meeting room / living room / workstation | 4-mic / XMOS DSP / AEC / ESP32S3 wake-word front end |
![]() | reSpeaker Clip | Pickup (wearable) | Worn on the body / up to 3 m / 360° | 2-mic PDM / SpeexDSP noise suppression / Bluetooth LE 5.3 + Wi-Fi |
![]() | ReSpeaker Flex Circular-4 | Pickup + playback (circular) | Built into devices, 360° / wake-word front end included | 4-mic / XMOS DSP / AEC / ESP32S3 / 10 W amplifier |
![]() | ReSpeaker Flex Linear-4 | Pickup + playback (linear) | Built into devices, 180° / wake-word front end included | 4-mic / XMOS DSP / AEC / ESP32S3 / 10 W amplifier |
The key variables for microphones are pickup distance and ambient noise. Key specs and recommended pickup + playback combinations are below.
Core Advantages
It depends on where the conversation model runs. With the on-device LLM (RK3588 with an RK1828 card, or J4012), recognition, conversation and speech synthesis all run on the device, and it works offline after images and models are downloaded on the first online start. With a cloud model, conversation needs the network; recognition and speech synthesis still run on the device.
It depends on the compute tier: the RK series and J3011 run a single channel; J4012 supports 2–3 channels; J5012 reaches 3+ — all figures are for the pure voice pipeline (recognition + synthesis) only. Once a local LLM is added, the model consumes the compute, so plan for a single channel. Multi-site deployments can also scale out with more devices.
Audio does not. Recognition and speech synthesis run on the device and recordings are not uploaded. With a cloud model, the recognized text is sent to the model endpoint you configure; with the on-device LLM, the text stays on the device too. Business-system integrations receive the recognized text.
Yes. The wake word can be any short Chinese or English phrase, applied at startup, with strict, balanced and sensitive settings. Voices come in machine, simulated and natural tiers; the natural-voice model on J4012 can clone a voice from reference audio.
1-year hardware warranty plus 1 year of technical support. Custom development work (own apps, deep customization) goes through the customization channel, scoped by the project team.
Yes. The Seeed devices used in this solution can be customized with your own logo, enclosure, packaging and pre-loaded firmware, and existing models can be adapted to add or drop interfaces and features. , or see Customization Service for scope and process.