Guides, references, and resources for DCYFR Labs tools and packages.
All DCYFR Labs packages are published to npm under the @dcyfr scope and available on GitHub.
Portable TypeScript AI agent framework with plugin architecture, multi-provider LLM support, delegation v2, built-in telemetry, and quality gates.
Command-line tools for AI-powered workflows. Config init, validation, and project scaffolding for teams building with @dcyfr/ai.
Retrieval-Augmented Generation library for semantic search and document grounding. Pluggable vector store adapters, plugs directly into the @dcyfr/ai plugin system.
Structured code generation with quality gates and Zod-validated output. Integrates with the @dcyfr/ai validation framework for guaranteed compliance.
Install the core framework and start building in minutes.
import { AgentRuntime, ValidationFramework, loadConfig } from '@dcyfr/ai';
// Zero-config setup — auto-detects .dcyfr.yaml
const config = await loadConfig();
// Connect to any provider
const runtime = new AgentRuntime({
provider: 'openai', // or 'anthropic', 'ollama', 'copilot'
model: 'gpt-4o',
});
const response = await runtime.chat({
messages: [{ role: 'user', content: 'Explain this codebase' }],
});
// Run quality gates before shipping
const validation = new ValidationFramework({ gates: config.validation.gates });
const report = await validation.validate({ projectRoot: '.', config: config.agents });
console.log(`Quality gates: ${report.valid ? '✓ PASS' : '✗ FAIL'}`);