VOXEDGE
On-device models
Acoustic and translation models tuned for mobile chipsets and battery budgets.
SOLUTIONS / SMARTPHONES
Every voice,
handled
on the phone.
Memory and live translation, running on the phone itself. No cloud, no per-call bill.
01 / SMARTPHONESON-DEVICE
VOICE → LANGUAGEVOXSENSE · VOXLINGO
01 / SCENARIOS
What people do with it.
01
Memory
Capture commitments and details from everyday conversations as local transcripts.
02
Translate calls
Recognize, translate and speak across languages during calls.
03
Talk face to face
Turn the phone into an interpreter for in-person conversations.
04
Assist
Connect context to reminders and timely suggestions.
02 / PRODUCTS INSIDE
Built from VoxNexus products.
VOXSENSE
Always-on intelligence
Low-power capture and on-device transcription as the foundation for memory and assistance.
VOXLINGO
Multilingual communication
Speech recognition, translation and synthesis in one on-device chain.
Estimated on a smartphone: 180 mAh incremental battery use over 12 hours, based on 15 mAh per hour. Reference setup: MTK9600, continuous speech playback; transcription once per minute.
03 / INTEGRATION
From microphone
to experience, on the phone.
The SDK provides acoustic front-end, VAD, ASR, MT and TTS. Understanding and assistance are built in your product. If the device has a low-power processor (e.g. SCP), VAD can run there around the clock in about 200 KiB of memory.
- YOU BUILD YOUR PRODUCT Captures the audio, feeds the SDK and uses the results. UNDERSTAND · ASSIST. Not in the SDK.
- VOXNEXUS SDK ON THE PHONE Acoustic front-end, VAD, ASR, MT and TTS. No cloud round trip.
- YOU BRING THE PHONE Runs on MediaTek Helio G81 and above.
- YOU BRING LOW-POWER PROCESSOR · OPTIONAL (E.G. SCP) Listens around the clock, runs VAD in about 200 KiB and hands speech segments to the SDK.
- YOU BRING MICROPHONES · AUDIO PATH Phone mics and call audio.
Illustrative architecture. Where processing runs is evaluated per device and configuration.
04 / EVALUATE
Bring your device. We bring the models.
Start with your phone’s chipset, audio path, languages and power budget. Then create an application, download the SDK and validate latency, power and memory on your own hardware.
Developer documentation ↗- Target device
- Smartphones
- Bring
- Chipset, optional low-power processor, microphone path
- Define
- Languages and scenarios
- Budget
- Power, memory, latency
- Integration model
- SDK + licensing