agno-agi/agno
SELF-MANAGING AI AGENT PLATFORM
Platform for building and running AI agents as durable services with API, MCP, and memory management. Automates company operations through agent teams and workflows.
// KEY FEATURES
- Run agents as services with API, MCP, and authentication
- Agent teams, workflows with conditions, loops, and parallel steps
- Cross-session memory, knowledge, tools, and human approval
- Deployment templates for Docker, AWS, GCP, Azure, Kubernetes
// DEEP DIVE: HOW IT WORKS
Agno is an AgentOS platform for running agents as durable services. Agents are built on SDKs (Agno, Claude Agent, Codex), combined into teams and workflows with routing, conditions, loops, and parallel steps. The platform provides REST API, MCP server for Claude/ChatGPT, and Slack interfaces. Stores conversations, memory, and activity in PostgreSQL, provides authentication, RBAC, execution resume, and human approval for sensitive actions.
Unlike orchestration-only frameworks, Agno includes a full service layer: scheduler, observability (traces, tool call inspection), sandboxes for Claude and Codex, and Company OS — a system for organizing work into products, projects, tasks, and recurring routines. Deployment templates for 9 clouds and Kubernetes. Coding agents run in isolated environments and can propose improvements via pull requests.
Requires own cloud or server for deployment (Docker, Railway, AWS, GCP, Azure, Fly.io, Render, Modal, Helm). Needs PostgreSQL for storage. Platform management and agent building assume coding agents (Claude/Codex) and skills from the agentos-railway repository. Data remains under user control.
// USE CASES
Best for: Engineers building production AI agents and automation platforms
- Deploy agent platform in your cloud with API and MCP for product
- Automate company routines: code review, task generation, monitoring
- Connect agents to Slack and Claude via MCP for team collaboration
// QUICK START
Clone https://github.com/agno-agi/agentos-railway into a folder called agent-platformCommands are taken from the repository README — check the current version on GitHub before running.
New repositories every 30 minutes
REDDYX AI scans GitHub 24/7 and ships the best AI/ML/Web3 projects to Telegram.
Join on Telegram