CortexON: Open source general AI agent, another alternative to Manus

What is CortexON?

CortexON is an open source general artificial intelligence agent system that can think and perform actions to automate routine tasks. It is inspired by advanced agent platforms such as Manus and OpenAI's DeepResearch, but unlike these platforms, CortexON is completely open source , making advanced AI technology accessible to everyone.

CortexON is developed by TheAgenticAI team, aiming to break the current situation where the latest AI breakthroughs are restricted by closed source code and paywalls, giving users data ownership and privacy control. The project is committed to building in an open environment and welcomes community contributions of ideas, feedback and code.

Core architecture and working principle

At its core, CortexON is a multi-agent AI system that integrates multiple specialized agents that dynamically collaborate to accomplish user-defined goals:

1. Web proxy : handles real-time Internet search, data retrieval, and Web interaction 

2. File Agent : Manages file operations, organization, data extraction and storage tasks 

4. Coding Agent : Generate, debug and optimize code snippets in various programming languages 

5. API proxy : seamlessly integrate with external services, APIs, and third-party software to expand automation capabilities 

These agents collaborate with each other and combine their unique capabilities to efficiently automate complex tasks.

Key Features

CortexON provides a range of powerful features:

-  Advanced, context-aware research automation

-  Dynamic multi-agent coordination

-Seamless  integration with third-party APIs and services

-Code  generation, debugging and execution

-Efficient  file and data management

-  Personalized and interactive task execution (such as travel planning, market analysis, educational content creation, and business intelligence)

Technology Stack

CortexON is built on the following technologies:

-Framework : PydanticAI multi-agent framework 

- Headless Browser : Browserbase (Web Proxy) 

-Search Engine : Google SERP 

- Logging and Observability : Pydantic Logfire 

- Backend : FastAPI 

-Front -end : React/TypeScript, TailwindCSS, Shadcn 

Quick Installation Guide

Environment variable configuration

To run CortexON, you need to create a .env file containing the following necessary variables :

Anthropic API

ANTHROPIC_MODEL_NAME=claude-3-7-sonnet-20250219ANTHROPIC_API_KEY=your_anthropic_api_key

Browserbase Configuration

BROWSERBASE_API_KEY=your_browserbase_api_keyBROWSERBASE_PROJECT_ID=your_browserbase_project_id

Google Custom Search

GOOGLE_API_KEY=your_google_api_keyGOOGLE_CX=your_google_cx_id

Logging

LOGFIRE_TOKEN=your_logfire_token

Vault Integration (Optional)

VITE_APP_API_BASE_URL=http://localhost:8000VITE_APP_VA_NAMESPACE=your_unique_namespace_idVA_TOKEN=your_vault_authentication_tokenVA_URL=your_vault_service_endpointVA_TTL=24hVA_TOKEN_REFRESH_SECONDS=43200

WebSocket

VITE_WEBSOCKET_URL=ws://localhost:8081/ws

Docker Setup

1.  Clone the CortexON repository:

git clone https://github.com/TheAgenticAI/CortexOn.gitcd CortexOn

2.  Set environment variables

3.  Build and run the Docker container:

docker-compose builddocker-compose up

As AI technology develops rapidly, CortexON will continue to evolve, integrating the latest AI research results to provide more powerful and intelligent automation capabilities. TheAgenticAI team is committed to maintaining the open source nature of the project to ensure the democratization and popularization of advanced AI technology.

CortexON represents the democratization of AI agent technology, making advanced AI automation capabilities accessible through open source. Whether researchers, developers or enterprise users, CortexON can be used to simplify daily tasks, optimize workflows, and explore the infinite possibilities of AI agents.