Skip to main content

​
Build Production-Ready AI Agents

Celesto scales your Agentic application and adds a security layer for MCP tools and agent-to-agent communication. Connect to external data using 100+ tools.
Celesto AI Platform

Quickstart

Build your first AI agent in 5 minutes

View on GitHub

Star us on GitHub and contribute

​
Why Celesto?

🚀 Fast Development

Build agents with simple Python APIs. No complex infrastructure setup required.

🔒 A2A Security

Built-in OAuth, authentication, and secure Agent-to-Agent communication.

🔌 100+ Integrations

Connect to databases, APIs, and external services via MCP protocol.

​
Quick Example

Build an AI Agent with tool access in just a few lines: 
from agentor.tools import WebSearchAPI
from agentor import Agentor

agent = Agentor(
    name="research_assistant",
    model="gpt-5",
    tools=[WebSearchAPI()]
)

# Run agent with natural language
result = agent.run("Find the latest AI research papers on arxiv")
print(result)
Check out our Quickstart Guide to build your first agent in minutes.

​
Core Features

Agentor - Build Agents

Open-source Python library for building and deploying agentic systems with security layer for Agent communication.

SuperAuth - Security Layer

Enterprise-grade security layer with OAuth 2.0, role-based access control, and secure agent-to-agent communication.

Tool Use & MCP Servers

Access 100+ prebuilt tools for real-time integrations with databases, APIs, and external services via MCP protocol.

Deploy Anywhere

Deploy agents with one command using the Celesto CLI with built-in auto-scaling, monitoring, and CI/CD support.

Need Help?

Join our Discord community to get help, share ideas, and connect with other developers.