AI-First logging library for intelligent agents
Log like you think. Universal, observable, and designed for the age of AI.

Traditional logging was built for traditional applications. But AI agents think differently - they reason, decide, reflect, and act across multiple contexts. Logent is the first logging library designed from the ground up for the age of intelligent agents.

AI agents need more than console.log:

• Context switching - Agents work across Node.js, edge runtimes, browsers, and tests
• Thought visibility - Need to trace reasoning chains, not just events
• Test-friendly - Logging shouldn't pollute test output
• Structured insights - LLMs need parseable, semantically rich logs

Logent provides:

• ✨ Universal - One logger, all runtimes (Node.js, Cloudflare Workers, Browser, Test)
• 🧠 AI-First - Designed for agent observability
• 🎯 Zero config - Auto-detects environment, sensible defaults
• 📊 Test-native - Silent by default in tests, inspectable logs
• 🚀 Tree-shakeable - Only bundle what you use

```bash npm install logent

pnpm add logent ```

```typescript import { createLogger } from 'logent';

const logger = createLogger();

logger.info('Agent started'); logger.warn('Low confidence in result'); logger.error('Tool call failed', { tool: 'search', reason: 'timeout' }); ```

See the Core Package README for complete documentation.

``` logent/ ├── packages/ │ └── logent/ # Core library ├── apps/ # Example applications (coming soon) ├── docs/ # Documentation └── operations/ # Internal tooling ```

Logent is built on three principles:

1. Universal - Write once, run anywhere
2. AI-First - Designed for intelligent agents
3. Simple - Zero config, sensible defaults

Inspired by the Monogent cognitive framework and built for the age of AI agents.

MIT © Deepractice

• PromptX - AI Agent development platform
• Monogent - AI cognitive framework

Logent - Logging for the age of agents.