Skip to main content
Version: 2.1.x

Appxiom Web SDK Installation

Detailed documentation is available at Appxiom Web SDK.

The official client-side JavaScript SDK for Appxiom. Use this package to initialize monitoring, track goals, report issues, and attach custom activity data from your web app.

Installation​

npm install @appxiom/web

Vite​

@appxiom/web includes a WebAssembly binary. During local development, exclude the package from Vite dependency optimization so Vite preserves the binary's source-relative URL:

import { defineConfig } from 'vite';

export default defineConfig({
  optimizeDeps: {
    exclude: ['@appxiom/web'],
  },
});

After adding this setting, clear Vite's optimized-dependency cache and restart the development server:

rm -rf node_modules/.vite
npm run dev -- --force

This setting affects Vite's development server only. It is safe to retain for production builds; Vite does not use dependency optimization when running vite build.

Production Deployment​

Run your framework's production build as usual. For Vite applications, this is typically:

npm run build

Vite emits the SDK's WebAssembly binary into the production output and rewrites the loader to its generated asset URL. Deploy the complete build output, including the generated .wasm asset, without renaming or omitting files.

Your CDN or web server must serve .wasm files with this response header:

Content-Type: application/wasm

Also ensure SPA fallback rules do not return index.html for .wasm asset requests. Receiving HTML instead of the binary produces errors such as expected magic word 00 61 73 6d.

For non-Vite applications, use the same deployment rules: the bundler must emit the package's .wasm asset, the asset must be deployed with the application, and the host must serve it with the WASM MIME type.

Connect to MCP Server​

Appxiom provides a Model Context Protocol (MCP) server that connects AI coding assistants and IDEs (such as Claude Code, Cursor, Windsurf, and VS Code) directly to Appxiom's real-time diagnostic engine. Once connected, your AI agents can inspect real-time application issues, occurrence telemetry, Quality Score (QS) metrics, and Goal Friction Impact (GFI) analytics directly within your editor.

To connect and configure MCP skills for your web application, see the MCP Skills Installation guide.