Note: Adapted from the official X Clawdi announcement at https://x.com/openclawdiai/status/2049883505187074150?s=46
If you’ve been using AI coding agents for any serious amount of work, you’ve probably noticed the same frustration. Claude Code on your laptop doesn’t remember what Codex did on another machine. Switch frameworks and you’re starting over. Every session is a blank slate, and all the context you built up just disappears. That’s not a tooling problem, it’s a fundamentally broken workflow, and it’s one that most developers have just quietly accepted as normal. Clawdi was built to fix it, and this latest update is the most complete version of that vision yet.
The idea is straightforward. Instead of your memory, files, API keys, and skills living inside a specific agent, they live in Clawdi. Every agent connects to that same environment. Switch from Claude Code to Codex to Hermes and nothing is lost, because the context was never tied to the agent in the first place. It all runs in a TEE-secured cloud, which means your keys and memory are private by default, not sitting on a shared server somewhere with no visibility into how they’re handled.
The Hermes Agent and What It Actually Does
The headline feature in this update is the Hermes Agent, now available to deploy in one click from the dashboard. What makes it different from a standard agent setup is that it builds on what it learned from previous tasks rather than resetting each time. It holds memory across sessions, picks up where Claude Code left off, and comes with over 200 tool integrations already configured so there’s no manual setup involved. If you’ve spent time getting an agent into a useful state only to lose that context when you close the session, Hermes is the direct answer to that.
A Dashboard That Actually Shows You Everything
The Clawdi Cloud dashboard was rebuilt around the idea of a single view for all your agents. Claude Code, Codex, Hermes, and OpenClaw sessions now appear side by side, with activity history, recent sessions, messages, memories, vault keys, and connectors all accessible from one place. Adding a new agent means pasting one prompt into your tool of choice and it configures itself and shows up in the dashboard automatically.
New additions include a built-in console with a terminal and file editor, an Agent Portraits page that gives each agent a shareable public profile, and a Connectors page with over 500 tools you can add in one click. Built-in skills now cover searching X posts, browsing live news, and checking Polymarket predictions out of the box, and messaging support spans 11 platforms including Telegram, Discord, Slack, and WhatsApp all managed from a single panel.
Pricing and How to Get Started
The free tier gives you access to either OpenClaw or Hermes running in hardware-secured infrastructure with support for 13 or more messaging apps. Pro at $29 a month unlocks both agents together along with a web terminal and custom ports. Max at $99 a month steps up to 4 vCPU, 8GB RAM, and 40GB storage. Enterprise covers SSO, audit logs, and a 99.9% uptime guarantee.
If you prefer to run it locally, the full version is available via npm install -g clawdi, MIT licensed and free to keep. It works with Claude Code, Codex, Cursor, OpenClaw, and Hermes. The whole thing takes under three minutes to get running and you can start for free at clawdi.ai.
Top comments (0)