@wavequery/conductor
v0.0.24
Published
Modular LLM orchestration framework
Downloads
343
Maintainers
Readme
Conductor | LLM Orchestration Framework
Build powerful AI applications with real-time visualization and dynamic orchestration. Conductor helps you create, monitor, and control complex LLM workflows through an intuitive chain and agent system.
Features
- Chain & Agent System: Build complex workflows combining multiple LLM tools
- Real-time Visualization: Monitor your LLM operations with interactive graphs
- Modular Architecture: Easily extend with custom tools and capabilities
- Multiple LLM Support: Works with OpenAI, Anthropic, and more
- Performance Tracking: Built-in metrics and logging
- Type Safety: Full TypeScript support
- Customizable UI: Flexible visualization options
🚀 Quick Start
npm install @wavequery/conductor
# or
yarn add @wavequery/conductor
Create Your First Chain
import { Chain, OpenAIProvider, VizServer } from '@wavequery/conductor';
// Initialize visualization
const vizServer = new VizServer(3000);
// Create your chain
const chain = new Chain({
name: "analysis-chain",
llmProvider: new OpenAIProvider({
apiKey: process.env.OPENAI_API_KEY
}),
tools: [analyzerTool, summarizerTool],
steps: [
{
name: "analyze",
tool: "analyzer",
input: { text: "Your text here" }
},
{
name: "summarize",
tool: "summarizer",
input: (prev) => ({
text: prev.analyze.text,
analysis: prev.analyze
})
}
]
});
// Execute and visualize
const result = await chain.runAgentLoop({});
// Visit http://localhost:3000 to see your workflow in action
Featured Examples
Market Analysis System
const analysisChain = new Chain({
name: "market-analysis",
tools: [technicalAnalyzer, sentimentAnalyzer, riskAnalyzer],
steps: [/* steps */]
});
const strategyAgent = new Agent({
name: "strategy-agent",
tools: [strategyGenerator, reportGenerator]
});
// Real-time visualization included!
📚 Documentation
Core Concepts
Guides
Example Projects
- 📝 Text Analysis
- Single Agent with Tool
- Real-time visualization
- 📊 Market Analysis
- Multi-Chain workflow
- Agent-based decision making
- ✍️ Content Generation
- Pipeline architecture
- Multiple tool coordination
Requirements
- Node.js >= 20.0.0
- OpenAI API key and/or Anthropic API key
- Environment variables setup (see Configuration Guide)
Contributing
We are extremely open to contributions, whether it be in the form of a new feature, improved infrastructure, or better documentation.
Please make a PR on github! Contribution guide will be added soon.
License
MIT © WaveQuery