A real-time WebSocket-based log streaming system with AI-powered analysis.

Drop-in replacement for the browser console API that routes all output over a WebSocket.

import { WebSocketLogger } from '@noconsole/logger';
const logger = new WebSocketLogger('ws://localhost:8080');

logger.log('Hello world!', { foo: 'bar' });
logger.warn('High memory', { usage: '92%' });
logger.error('Unhandled rejection', err);
// Sends: { type, args, timestamp, id }

• Full Console interface (log, warn, error, info, debug, time/timeEnd, count, table, group, …)
• WebSocket | null — caller owns the socket lifecycle
• Silent no-op when socket is null or not OPEN
• Full TypeScript support

All-in-one viewer: a WebSocket broadcast server plus the React web client in a single package.

Minimal WebSocket server that accepts connections from @noconsole/logger instances and broadcasts each message to all connected viewer clients.

npm run server --workspace=packages/streamer
# or
node packages/streamer/server/index.js

Listens on ws://localhost:8080 by default (override with PORT=…).

Real-time log viewer served at http://localhost:5173 in dev mode.

• Virtualized list (@tanstack/react-virtual) — handles thousands of entries without degradation
• Level filters: All / Log / Info / Warn / Error / Debug
• Search: real-time substring match with inline highlight; case-sensitivity toggle
• Pause/Resume: buffers up to 10k logs while paused, flushes on resume
• Auto-scroll with smart manual override
• AI Analysis panel: streams logs to a local Ollama instance, response parsed into Errors / Warnings / Performance / Recommendations
• Export logs as JSON

noconsole/
├── packages/
│   ├── logger/          # @noconsole/logger — WebSocket logger library
│   └── streamer/        # @noconsole/streamer — WS server + React viewer
│       └── server/      #   broadcast server (Node.js)
└── package.json         # workspace root

# 1. Install all dependencies
npm install

# 2. Build all packages
npm run build

# 3a. Start the broadcast server
npm run server

# 3b. In another terminal, start the dev viewer (client + server together)
npm run dev

npm run dev inside packages/streamer starts both the WS server and the Vite dev server concurrently via concurrently.

interface LogMessage {
  type: 'log' | 'warn' | 'error' | 'info' | 'debug' | ...;
  args: any[];
  timestamp: number;
  id: string;
}

• WebSocketLogger.ts — implements the Console interface
• index.ts — public API
• index.test.ts — Jest tests

• server/index.js — Node.js WebSocket broadcast server
• src/ — React + TypeScript viewer
  • Hooks: useWebSocket (connection + log state), useLLM (Ollama)
  • Components: Header, FilterBar, LogViewer, LogRow, LLMPanel

# Root
npm install           # install all workspace dependencies
npm run build         # build all packages
npm run dev           # start streamer dev server (WS + Vite)
npm run server        # start WS broadcast server only
npm test              # run logger tests

# packages/logger
npm run build
npm test

# packages/streamer
npm run dev           # WS server + Vite (concurrent)
npm run build         # production build
npm run server:start  # WS server only

1. Install Ollama and pull a model: ollama pull llama3.2
2. Start Ollama (http://localhost:11434)
3. Click AI Analysis in the viewer
4. Select your model and click Analyze Logs

MIT