Skip to main content
OmniCoreAgent Light

​
The Cognitive Runtime for ASI

OmniCoreAgent is a production-grade framework for building autonomous AI agents that persist, reason, and scale. Unlike other frameworks that treat memory as an afterthought, OmniCoreAgent puts State Management at the center. 
pip install omnicoreagent

​
Get Started

Quick Start

Build your first agent in under 30 seconds

Installation

Install with pip, uv, or from source

​
Core Capabilities

Multi-Tier Memory

5 backends (Redis, PostgreSQL, MongoDB, SQLite, in-memory) with hot-swap at runtime

MCP Client

Built-in Model Context Protocol — stdio, SSE, and Streamable HTTP transports

Context Engineering

Dual-layer context management so agents never hit token limits

DeepAgent

Multi-agent orchestration with the RPI+ workflow for complex research tasks

40+ Community Tools

Search, code execution, APIs, databases, and more — ready to use

OmniServe

Production API server with WebSocket streaming and multi-tenancy

​
See It In Action

import asyncio
from omnicoreagent import OmniCoreAgent

async def main():
    agent = OmniCoreAgent(
        name="my_agent",
        system_instruction="You are a helpful assistant.",
        model_config={"provider": "openai", "model": "gpt-4o"},
        mcp_tools=[{
            "name": "filesystem",
            "transport_type": "stdio",
            "command": "npx",
            "args": ["-y", "@modelcontextprotocol/server-filesystem", "/tmp"]
        }]
    )

    result = await agent.run("Summarize all .md files in /tmp", session_id="user_1")
    print(result["response"])

    await agent.cleanup()

asyncio.run(main())

​
Explore the Docs

Architecture

Layered design — ReAct, Storage, and Connectivity

Event System

Real-time streaming for every tool call, thought, and message

Workflows

Sequential, parallel, and router patterns for multi-agent tasks

Model Support

9+ providers — OpenAI, Anthropic, Gemini, Groq, DeepSeek, and more

Observability

Opik tracing, per-request metrics, and production monitoring

Configuration

Environment variables, agent settings, and model configuration