Fragola Agentic SDK

Build AI Agents Your Way

Fragola is an event‑driven SDK for building AI‑first software and custom agents. It focuses on composable primitives, predictable state, and production‑ready patterns for orchestrating AI tools and workflows.

Status: Under construction 🚧 – first public beta releases coming soon. 🧪

• 🌐 Website: https://fragola.ai
• 📚 Docs: https://docs.fragola.ai

Key Features

• [x] ⚙️ Core primitives: Define agents, tools, events, and state transitions as first‑class building blocks instead of ad‑hoc prompts and handlers.
• [x] 📡 Event‑driven model: Drive your agents through explicit events (user messages, tool results, system triggers) for easier debugging and observability.
• [x] 🧩 Lightweight and extensible: Bring only what you need; extend primitives to match your product’s architecture and domain.
• [x] 🏭 Production‑ready focus: Built for long‑running agents, real‑world error handling, and integration into existing backends and UIs.

Built‑in Hook Presets 🔌

Fragola ships with ready‑made “hook presets” for common patterns:

• mcpClient: Connect to remote or local MCP servers and expose their tools to your agents.
• orchestration: Coordinate multiple specialized agents that can communicate and collaborate on a task.
• guardrail: Validate user messages against unwanted or unsafe content and reject them with a clear reason.
• fileSystemSave: Persist conversations and state to the filesystem on Node.js‑like runtimes.

Getting Started 📦

The SDK is not yet publicly released. We’re finalizing APIs and examples and will publish:

• An installable package: @fragola-ai/agentic-sdk-core
• Full quickstart guides and recipes
• Integration examples for popular runtimes and frameworks

Quickstart Example

Install the SDK:

npm install @fragola-ai/agentic-sdk-core

Create a simple agent that responds to user messages:

import { Fragola } from "@fragola-ai/agentic-sdk-core";

const fragola = new Fragola({
	apiKey: process.env.OPENAI_API_KEY, // Can also be set in your environment
	model: "gpt-4o-mini"
});

const agent = fragola.agent({
	name: "QuickstartAgent",
	description: "My first Fragola agent",
	instructions: "You are a helpful and concise assistant."
});

const state = await agent.userMessage({
	content: "Hello! What can you do?"
});

console.log(state.messages); // Full conversation history