@cohesiumai/providers-webnn

WebNN provider for browser-ai — Hardware-accelerated inference via Web Neural Network API.

Installation

pnpm add @cohesiumai/providers-webnn

Usage

import { createWebNNProvider } from '@cohesiumai/providers-webnn';

const provider = createWebNNProvider();

// Check availability
const result = await provider.detect();
if (result.available) {
  console.log('WebNN is available');
}

Browser Support

| Browser | Support | |---------|---------| | Chrome 127+ | ✅ Full | | Edge 127+ | ✅ Full | | Chrome <127 | ⚠️ Flag required | | Firefox | ❌ Not supported | | Safari | ❌ Not supported |

Features

• Hardware Acceleration — Uses GPU/NPU when available
• Native Performance — Direct access to ML accelerators
• Automatic Fallback — Works with browser-ai's provider selection

API

createWebNNProvider(options?)

Creates a new WebNN provider instance.

interface WebNNProviderOptions {
  devicePreference?: 'gpu' | 'cpu' | 'npu';
}

Provider Interface

• detect() — Check if WebNN is available
• init(config, modelSpec) — Initialize with model
• generate(params) — Run inference
• abort() — Cancel current generation
• teardown() — Cleanup resources

V0.2+

This provider was introduced in browser-ai V0.2.

License

MIT © 2026