Fynq

Orchestrate Intelligence.

The package manager for the agentic age.
Discover, install, and run autonomous agents with the simplicity of a single command.
Built for the next generation of AI engineering.

curl -fsSL https://fynq.sh/install | sh

Features  •  Quick Start  •  Registry  •  Architecture

---

✨ The Fynq Experience

Fynq reimagines how we interact with AI. It is not just a tool; it is a protocol for autonomous execution. Just as npm unlocked the power of JavaScript modules, fynq unlocks the power of composable AI agents.

Why Fynq?

• ⚡ Native Performance: A lightweight CLI written in Python, optimized for speed.
• 🔒 Secure Sandbox: Agents run with least-privilege permissions. You decide if they can access the web or your files.
• 🌍 Universal Registry: Publish your intelligence once, run it anywhere. Versioned, signed, and immutable.
• 🔌 Model Agnostic: Bring your own LLM. Use GPT-4 for reasoning, Haiku for speed, or a local Llama 3 for privacy.

---

🚀 Quick Start

From zero to autonomous in 10 seconds.

1. The Summoning

Install the universal runtime.

curl -fsSL https://fynq.sh/install | sh

2. First Contact

Summon a researcher agent to analyze a topic for you.

# Analyze a YouTube video and write a report
fynq run @fynq/youtube --task "Analyze the MFE paradigm in https://youtu.be/..." 

---

🛠️ Building Intelligence

Become a creator. Build agents that act on your behalf.

1. 🏗️ Initialize

Create a new workspace with a standardized structure.

fynq init my-agent
cd my-agent

2. 🧠 Implement

Write your agent logic in main.py. The fynq SDK handles the heavy lifting.

# main.py
import fynq
from fynq import Agent

def main():
    # The runtime injects the user's task automatically
    task = fynq.get_task() 
    
    # Use the 'browser' tool capability
    content = fynq.tools.browser.visit("https://example.com")
    
    # Analyze with the configured LLM
    summary = fynq.llm.chat(f"Summarize this: {content}")
    
    print(summary)

3. 📦 Publish

Share your creation with the world.

fynq publish
# 🚀 Successfully published @your-name/my-agent v0.1.0

---

🏗️ Architecture

The Fynq ecosystem is composed of three harmonious layers.

graph TD
    subgraph "💻 User Space"
        CLI[("Fynq CLI")]
        Agent[("Agent Runtime")]
        Sandbox[("Secure Sandbox")]
    end

    subgraph "☁️ Cloud Layer"
        Registry[("Global Registry")]
        Auth[("Authentication")]
        Storage[("Artifact Storage")]
    end

    CLI -->|Pull Agent| Registry
    CLI -->|Verify Signature| Auth
    CLI -->|Run| Agent
    Agent -->|Execute| Sandbox
    Sandbox -->|Restricted Access| Internet((Internet))

Loading

---

📦 The Registry

A curated collection of verified agents.

Package	Description	Version
@fynq/researcher	Deep web researcher. Browses, reads, cites sources.	v1.2.0
@fynq/coder	Full-stack software engineer. Edits files, runs tests.	v0.8.5
@fynq/youtube	Watch & analyze videos. Extracts timestamps & summaries.	v1.0.1
@fynq/data	SQL & CSV analyst. Visualizes data trends.	v0.9.0

---

Join the Revolution

 

Crafted with 🖤 by the Fynq Team.
© 2026 Fynq Inc. All rights reserved.