@overmorrow-labs/bud v1.3.0

Bud

Bud is a client-side framework designed to abstract away the complexity of using LangChain and LangGraph. It provides a lightweight and modular agent orchestration layer that connects to LLMs via OpenAI-compatible APIs or local Ollama setups. Bud makes it easy to spin up, manage, and interact with multiple AI agents for different features in your application.

✨ Features

• Agent Orchestration over LangChain / LangGraph
• Plug-and-play support for OpenAI APIs and Ollama via connection string/config
• React components for chat-based experiences (Assistant UI)
• Instance-based multi-agent support
• Context/configuration-based agent initialization
• Simple API to build custom UX on top of AI agents

🧠 Scope

• Abstracts LangChain / LangGraph logic
• Provides a client-side Agent Orchestration Service
• Exposes reusable React components like:
  • FloatingAssistantBud
  • AssistantBud
• Enables creating multiple agent instances using unique instanceIds
• Supports custom context/tooling configuration via exposed interfaces
• Ideal for apps needing multiple AI agents across different features
• Sits on the client side, enabling seamless interaction with models through APIs or local endpoints

🔁 Developer Flow

1. Install Bud

   npm install @overmorrow-labs/bud
   yarn add @overmorrow-labs/bud

2. Create an Agent Instance

   import { AgentOrchestrationService, createConfig } from '@your-scope/bud';
   
   const config = createConfig({
     context: "You're a helpful assistant...",
     tools: [/* Optional tools */],
     model: 'openai:gpt-4',
     connection: {
       apiKey: 'sk-...',
       endpoint: 'https://api.openai.com/v1'
     }
   });
   
   AgentOrchestrationService.createInstance('my-product-agent', config);

3. Access Agent Anywhere

   const instance = AgentOrchestrationService.getInstance('my-product-agent');
   await instance.ask('How can I help you today?');

4. Use Built-in UI Components

   import { AssistantBud, FloatingAssistantBud } from '@your-scope/bud/ui';
   
   <AssistantBud instanceId="my-product-agent" />
   // or
   <FloatingAssistantBud instanceId="my-product-agent" />

5. Custom Config Builder

   Bud exposes interfaces to build the context config programmatically. This allows setting up agents with relevant base instructions, tools, and MCPs to perform specific tasks effectively.

🎯 Use Cases

• Embed multiple AI agents in different sections of your app (e.g. support, onboarding, documentation, etc.)
• Provide contextual chat-based features with minimal integration
• Extend agent logic via LangChain tooling while keeping the interface lean

🧩 TL;DR

Bud is your agent superlayer. It:

• Manages agent lifecycles
• Bridges frontend with LLMs
• Gives you beautiful React components
• Allows unlimited agent instances, each tailored for your feature
• Helps you do wonders ✨