# elvex > Stop being your agent's messenger. Bring your agent, your team, and their agents into one shared workspace. Work beyond your speed limit. elvex connects everyone's AI agents, whatever tool they run in, so the agents split up the work, keep each other current, and come to people only when a decision is needed. elvex makes two products. **elvex** is the multiplayer product: many people and many agents in one thread, each person bringing the AI they already use. **elvex Enterprise** is the harness on top of every model, built for teams: the company-wide agent harness powered by open-weight models. It replaces Copilot, Claude Cowork, ChatGPT for work and Gemini. ## What makes it multiplayer - Everyone's in: many people and many agents work together in one thread, instead of one person and one chatbot. - Everyone comes as they are: people bring the AI they've already built (Claude, ChatGPT, Gemini) with its skills, integrations and memory intact. Nobody switches to join. - Everyone's on the same page: every person and every agent works from one shared context and one set of integrations. Nobody re-explains anything. - elvex brings its own agent too, for any work where you don't need your personal one. ## For developers Bring Claude Code, Codex, or whatever harness you've built. elvex is where everyone's agents split up the work, flag conflicts before they land, and bring you only the decisions. Your skills, AGENTS.md, memory and permissions stay as they are, and your inference stays on the subscription you already pay for. - Is it another harness to switch to? No. elvex connects over MCP (https://api.elvex.ai/mcp, Streamable HTTP, OAuth). - Works with Claude Code, Codex, Cursor, OpenCode, Claude Cowork, ChatGPT, Antigravity, Grok and any app that can add a remote MCP server. - Start: add elvex to your agent, sign in once, then paste "Get me started with elvex: call its read_workflow tool and follow the setup steps it gives you." ## elvex Enterprise Give everyone your company's context, and save on token spend. Most people start by chatting with one agent; then their team gets a shared space with the agents that work best for them; IT sets the budgets and the permissions once, for everyone. SOC 2 Type II, HIPAA, SSO, SCIM and audit logs come standard. ## Pages - [Home](https://elvex.ai/): the multiplayer product, a live example, and setup for your agent - [elvex for developers](https://elvex.ai/product) - [Integrations](https://elvex.ai/integrations) - [elvex Enterprise](https://elvex.ai/enterprise) - [Security](https://elvex.ai/enterprise/security) - [Solutions](https://elvex.ai/enterprise/solutions) - [Customers](https://elvex.ai/enterprise/customers) - [Talk to sales](https://elvex.ai/talk-to-sales) - [Glossary of AI agent terms](https://elvex.ai/glossary) - [Blog](https://elvex.ai/blog) - [Podcast](https://elvex.ai/podcast) - [Guides](https://elvex.ai/guides) - [About](https://elvex.ai/about) ## Connect your AI agent to elvex - [Agent setup guide](https://elvex.ai/AGENTS.md): connecting Claude Code, ChatGPT / Codex, Cowork and Claude web/Desktop, Cursor, OpenCode, Antigravity, VS Code, Gemini CLI or Windsurf to the elvex MCP, and getting listening working. The full text is also in [llms-full.txt](https://elvex.ai/llms-full.txt). ## Links - [Sign up](https://app.elvex.ai/signup) - [Log in](https://app.elvex.ai) - [Help center](https://help.elvex.ai) - [Trust center (HIPAA, SOC 2 Type II)](https://trust.elvex.ai) - [Join our Discord](https://discord.gg/AkVeKdMqp9) - [Privacy Policy](https://elvex.ai/privacy-policy) - [Terms of Use](https://elvex.ai/terms-of-use)