On-Device AI — No Cloud, No Internet, No Compromise

SLM360 LensNano
869KB of AI that replaces your cloud.

The complete on-device voice + vision engine for smart glasses. 15 intents. Entity extraction. Hindi + English. Sub-10ms. Fits on any ARM chip.

1.67M
Parameters
869KB
Model Size
<10ms
Latency
100%
Benchmark
Try Live Demo Talk to Us
How It Works
Two-stage pipeline. Always on. Always private.

Stage 1 runs continuously at ultra-low power. Stage 2 activates only when needed. Total on-device footprint: 55MB.

Stage 1 — Always On (17MB, ~5mW)
Microphone
Always listening
360 VAD
2MB
360 Wake
15MB
Stage 2 — On Demand (43MB)
360 STT
42MB
SLM360 NLU
869KB
Action
Instant
Capabilities
Everything on-device. Everything instant.

15 Voice Intents

Navigate, call, message, capture, music, fitness, weather, alarms, shopping, translate, notifications, focus mode, identify, read text, assistant.

Entity Extraction

"Call Mom" extracts contact="Mom". "Navigate to Khan Market" extracts location. Contacts, times, durations, languages, frame types.

360 OCR Vision

On-device neural OCR engine. Read signs, menus, prescriptions, labels through the glasses camera. No cloud processing.

Hindi + English

Native Hinglish code-mixing. "Mujhe directions do", "papa ko call karo" classified correctly. 99 languages supported.

100% Private

Every computation on-device. Zero data leaves the glasses. No cloud. No internet. Works in airplane mode, rural areas, everywhere.

OEM-Ready SDK

Configurable wake word per brand. Rust SDK, ARM-native, ONNX export for any SoC. Drop-in integration.

Comparison
Every competitor needs the cloud. We don't.
FeatureSLM360 LensNanoRay-Ban MetaXrealGoogle Glass
Latency<10ms2-3s2-5s1-3s
Works OfflineYesNoNoNo
HindiYesNoNoLimited
PrivacyOn-deviceCloudCloudCloud
Total Size55MBCloudCloudCloud
Cost/Query₹0₹0.80-4₹1.60₹0.80
OCROn-deviceCloudNoCloud
Entity ExtractionYesLimitedNoLimited
Wake WordConfigurable"Hey Meta"Fixed"OK Glass"
Languages99EnglishEnglish4
Live Demo
Try it. Right now.

Type a command or use the mic. The model runs on-device in a single binary.

360 OCR Vision
See text. Understand text. Act on text.

The glasses camera reads text in the real world — signs, menus, prescriptions, labels — then classifies what the user wants to do with it. Entirely on-device.

Tap a scenario to simulate the glasses camera
Menu Board
Coffee shop menu with prices
Road Sign
Street sign with directions
Medicine Label
Prescription with dosage info
Hindi Sign
Bilingual store signage
Or upload your own image
Cost Savings
Zero marginal cost. Forever.

The model runs on the chip. No API calls. No per-query fees.

1,000,000
Users × 50 commands/day
Cloud Cost / Year
₹90 Cr
LensNano / Year
₹0
Under The Hood
Purpose-built for the edge.
Architecture
192-dim, 4 Layers, 6 GQA Heads
Activations
SwiGLU + RoPE + RMSNorm
Parameters
1,676,751
Quantized Size
869 KB (INT4)
Inference
<10ms on CPU
Benchmark
100% (52/52)
Framework
360 Engine (Rust, zero-GC)
GPU Training
360 Tensor (Metal / CUDA)
Tokenizer
360 BPE (2K vocab)
Edge Deploy
360 Edge (ARM / RISC-V)
OCR Engine
360 Vision (neural OCR)
Platforms
ARM, x86, WASM, RISC-V
STT Engine
360 Speech (99 languages)
VAD
360 VAD (always-on, 2MB)
Wake Word
360 Wake (configurable)
License
Proprietary SDK
The Engine
Powered by SLM360X

LensNano is built on SLM360X — our most advanced on-device AI engine. A ground-up rewrite of the entire inference stack: custom transformer architecture, GPU-accelerated training, proprietary tokenization, neural OCR, and edge deployment — all in a single unified framework.

SLM360
Foundation — Gen 1

Where it started. Hand-crafted tensor operations, byte-level tokenization, custom backward pass. Built from scratch in Rust with zero external dependencies. Proved that sub-millisecond on-device NLU was possible.

Custom Tensors SGD Training CPU Only
Current
SLM360X
Next-Gen — Gen 2

Everything rebuilt from the ground up. GPU-accelerated tensor engine with Metal and CUDA support. Proprietary BPE tokenizer. Neural OCR for vision. ONNX-based edge deployment for any ARM or RISC-V SoC. The engine that powers LensNano.

GPU Accelerated AdamW + Autograd BPE Tokenizer Neural OCR Edge ONNX
SLM360X Model Family
SLM360 Nano
8M params · 4MB · General NLU
SLM360 LensNano
1.67M params · 869KB · Smart Glasses
SLM360 Base
120M params · 60MB · Conversational AI