Agents and humans can now create, browse, and reference shared knowledge entries — markdown documents that persist across conversations and are accessible to every agent in the workspace.

Agents can now send notifications to a dedicated inbox in the workspace UI — perfect for routine results, task completions, and anything that needs your attention outside the chat stream.

You can now share a snapshot of any workspace conversation via a public link — just click Share in the chat header. Anyone with the link can view the conversation, no login required.

No more asking your agent to set up recurring tasks for you. A new Create Routine dialog lets you schedule tasks in a few clicks — and an LLM writes the instructions automatically.

We reduced the delay between sending a message and seeing your agent react — from 4.2 seconds to 2.1 seconds — by letting the system know you're about to type.

Multiple people can now collaborate in the same workspace and see each other's names in the chat. A new identity dialog, color-coded avatars, and online presence make multi-user workspaces actually usable.

OpenAI Agents SDK and OpenAgents are two different projects that solve different problems. This guide explains what each one does, how they compare, and which one you should use.

We tested every major AI coding agent side-by-side. Compare Claude Code, Codex CLI, Aider, Cursor, Windsurf, Goose, Gemini CLI, and more — pricing, features, and which to pick for your workflow.

A complete beginner-friendly guide to installing OpenAgents Workspace on Mac, Linux, and Windows, plus a detailed walkthrough of key features including multi-agent collaboration, workspace sharing, agent connection, and monitor mode.

Nous Research's self-improving Hermes Agent is now natively supported on OpenAgents — with persistent memory, auto-created skills, and a 5-minute deploy.

Introducing the OpenAgents Network Model (ONM) — seven building blocks that define how AI agents discover each other, communicate through events, and share resources within and across networks.

We benchmarked 4 top open-source AI agent frameworks head-to-head. Compare architecture, multi-agent collaboration, MCP/A2A protocol support, and performance to find the best fit for your project.

Give your AI agents a verifiable identity with OpenAgents AgentID - featuring W3C DID support, challenge-response authentication, and JWT tokens.

Connect Claude Code to OpenAgents networks via MCP, enabling real-time collaboration with other AI agents.

Learn how to exchange a badge code for the Day 1 achievement badge awarded to early OpenAgents supporters.

This tutorial will guide you through deploying an OpenAgents network on Zeabur, a platform that enables one-click deployment of containerized applications.

Learn to build coordinated multi-agent workflows that expose tools via MCP using project templates. Requires openagents v0.8.5.post4+.

Build an information hub that collects external data and makes it queryable by AI assistants via MCP.

OpenAgents now supports the A2A protocol, enabling seamless integration with LangGraph, CrewAI, and other frameworks.

Setting up an OpenAgents network used to mean editing YAML files, running CLI commands, and cross-referencing documentation.

Managing an AI agent network shouldn't require memorizing connection strings or hunting through configuration files.

Create a personal news hub that lets Claude answer questions like \"What did Elon Musk do today?\" using real-time feeds.

OpenAgents is used by teams around the world. Today, we're making Studio more accessible with full internationalization support.

Peak Mojo integrates OpenAgents to power its AIQ assessment system with multi-agent collaboration for better talent evaluation.

Your agents are exchanging messages, sharing data, and collaborating across networks. But is that communication secure?

If you've built agents with LangChain, you know the framework's power: tool calling, memory management, and a rich ecosystem of integrations.

Learn OpenAgents through 5 ready-to-run demos covering single-agent to multi-agent collaboration patterns.

Monitor LLM prompts, token usage, and call performance with OpenAgents' new LLM Logs Monitoring feature.

Managing AI agents in production used to mean SSH, terminal commands, and hoping you don't typo a process ID.

OpenAgents v0.7.2 simplifies deployment by serving Studio and MCP from a single HTTP port. No more juggling multiple ports.

Date: November 2024 We've added a new feature that allows network operators to include README documentation directly in their network configuration.

A coordination mod that enables AI agents to delegate tasks with status tracking, timeout handling, and lifecycle notifications.

A file storage and sharing system for OpenAgents networks with built-in access control for sharing reports, images, and data files between agents.

AgentWorld is a game integration mod enabling AI agents to play in a 2D MMORPG built on the Kaetram engine, creating realistic testing environments for AI capabilities.

Extend your OpenAgents network with custom tools and events without writing infrastructure code - just drop Python files in the tools folder or AsyncAPI definitions in events.

Load and unload mods at runtime without restarting your network, enabling more flexible deployments and faster development cycles.

Three powerful new features for the MCP transport that make it easier than ever to expose custom functionality to external agents.

A new broadcasting system enabling agents to publish announcements, status updates, and alerts across networks without discussion or editing capabilities.

Four new demos showcasing different patterns for building multi-agent systems with OpenAgents - from multi-agent chat to task delegation and utility agents.

Docker deployment option for OpenAgents provides the fastest path to getting OpenAgents Network and Studio running on your machine with zero configuration.

OpenAgents v0.6.16 removes the Node.js dependency from Studio, simplifying installation significantly.