3 packs, 24 hours
700 mAh each, roughly 8 hours per pack, swapped without powering down.
Three-button control
Power and volume up/down. Found by feel, no touch surface to misfire.
Any language
Language ability comes from your models, not the hardware. It speaks whatever your stack speaks.
Wi-Fi + BLE
Streams over the site's own network on a loss-tolerant UDP protocol.
Your AI
Configurable endpoint, device behaviour tuned per use case from the console. Built to embed in the voice AI project you already have.
Rechargeable dock
The dock carries its own charge, so packs top up away from mains.
Your voice AI, worn on one ear
QuickComm Earpiece v3 · SKU QC-EP-V3
Line total
$220
Volume pricing
A development kit for engineers, developers and builders. Ships pre-flashed and streams to the endpoint you configure from the console: point it at your own backend, run your own stack, and it slots into the voice AI project you already have, with no per-channel fees and no lock-in.
The same path on all three devices. Only where the microphone sits changes.
Press to talk. Nothing is captured until then.
16 kHz mono over Wi-Fi. Silence never streams.
Clean, framed utterances arrive at your endpoint.
Your agent, your rules, on your infrastructure.
Answers spoken back, records kept, in real time.
HOT-SWAP POWER: ships with three 700 mAh detachable battery packs. Each pack runs about 8 hours, so a full set covers a 24-hour rotation without the device ever leaving the ear.
RECHARGEABLE DOCK IN THE BOX: the dock holds its own charge, so it tops up spare packs away from a wall socket. Swapping takes seconds and needs no tools.
THREE PHYSICAL BUTTONS: power and volume up/down, laid out so they are found by feel. No touch surface to misfire when it is wet, cold or gloved.
BUILD YOUR OWN VOICE AI ON IT: this is off-the-shelf hardware you can make your own. Point it at your own backend, run your own models, and tune how the device behaves from the console. It is built to embed in the project you already have.
MADE TO BE EMBEDDED: streams push-to-talk-framed 16 kHz audio to the endpoint you configure, so it drops straight into the voice AI project you already have.
CAPTURE YOUR AI CAN TRUST: push-to-talk framing means silence is never streamed, and a pre-roll buffer means the first syllable of a transmission is never clipped.
ONE DEVICE, ANY JOB: team push-to-talk, a meeting note taker, a personal assistant you can just talk to. Whichever of those your project is, the hardware is the same.
MANAGED AS A FLEET: QR-code provisioning, health telemetry, remote commands and over-the-air firmware updates.
| Battery type | Detachable rechargeable LiPo |
|---|---|
| Battery capacity | 700 mAh per pack |
| Runtime per pack | Approx. 8 hours |
| Packs included | 3 (approx. 24 hours combined) |
| Hot-swap | Yes. Packs change without powering the unit down |
| Charging | USB-C fast charging |
| Charging dock | Included, and rechargeable. Holds its own charge to top up packs away from mains |
| Buttons | Three physical buttons: power, volume up, volume down |
|---|---|
| Status | RGB LED indicator |
| Audio out | Integrated speaker |
| Microphone | Digital MEMS |
| Form factor | Over-the-ear hook, single ear |
|---|---|
| Fit | Left or right ear, ambidextrous hook |
| Housing | Matte soft-touch polymer |
| Wi-Fi | 802.11 b/g/n, 2.4 GHz, integrated antenna |
|---|---|
| Bluetooth | Bluetooth 5.2 (BLE) |
| Stream transport | QuickComm udp-v1 push-to-talk wire protocol |
| Packet resilience | Self-describing packets, per-boot session keys, loss tolerant |
| Provisioning | QR-code pairing with MAC-based device identity |
| Microphone | Digital MEMS microphone |
|---|---|
| Audio codec | 24-bit ADC/DAC |
| Capture format | 16 kHz PCM, mono |
| Gating | Push-to-talk framed. Silence is never streamed |
| Pre-roll buffer | Yes. The first syllable of a transmission is never clipped |
| Edge processing | On-device voice activity detection and wake handling |
| Flash memory | 8 MB |
|---|---|
| Firmware updates | Over-the-air, staged rollouts from the fleet console |
| Remote management | Remote commands, health telemetry, offline detection |
| Backend | Yours: the stream endpoint is configurable from the console |
|---|---|
| Firmware | Ships pre-flashed; device behaviour is configured per use case from the console |
| Voice AI stack | Bring your own STT, LLM, wake word and TTS |
| On-device logic | 8 MB flash with on-device voice activity detection built in |
| Hardware platform | Standard, well-documented parts, with no proprietary ASIC to design around |
| Warranty | 12 months limited hardware warranty |
|---|---|
| Designed & engineered | Dubai, UAE (in5 Tech, Dubai Design District) |
| Manufacturer | Quickgen Technologies |
| IP status | Patent pending |
Custom hardware for any voice AI application
Point the Earpiece at your own backend and tune it for your use case from the console.
Every unit ships pre-flashed with QuickComm firmware, ready out of the box. Nothing is ever reflashed in the field.
Pick how the device behaves for your use case, then set the stream endpoint to your own infrastructure.
The audio goes where you send it. Bring your own speech-to-text, models, wake word and text-to-speech.
On-device intelligence. 8 MB of flash and voice activity detection keep capture clean before a byte leaves the device.
Always current. Staged over-the-air updates, with fleet health telemetry behind them.
No provider lock-in. Nothing about the hardware ties you to a particular AI provider.
Answered by the QuickComm engineering team.
Something not covered here? Ask us directly. We answer hardware questions before you buy, not after.
QC-TT-V3
$90
per unit
Rechargeable desk or wall unit for room capture and note taking, with an on-device status display and a 1000 mAh battery so it keeps working away from a socket.
Best for: Control rooms, front desks, briefings and note taking
QC-BN-V3
$125
per unit
Plugs into the audio jack of any device: radio, handset, intercom, PA or console. Two-way mic and speaker, streaming straight to your backend to feed whatever you run there.
Best for: Any audio source you want feeding your voice AI
