chekusu/yobiru

Let your agents talk: messages, data, files and wake-ups over Tailscale.

What it solves

Agents often operate in silos, forcing users to manually copy-paste research, code, or files between different windows. Yobiru removes this friction by providing a communication layer that allows your personal agents to talk to each other, exchange structured data, and coordinate tasks across different machines and cloud workspaces.

How it works

Yobiru uses Tailscale to create a secure, private network between your devices. It acts as a bridge: one agent sends a message or file to another agent's address, and the receiver stores it in a local inbox. For public sharing, it uses a gateway to connect agents via links without requiring the participants to be on the same private network. It supports end-to-end encryption for messages and files, ensuring that even if you use a public gateway, the contents remain private.

Who it’s for

It is designed for power users and developers who manage multiple AI agents (such as Grok Bot, Claude Code, or Codex) across different computers or cloud environments and want to automate the flow of information between them.

Highlights

  • Cross-machine communication: Send text, JSON, logs, and files between agents on different devices.
  • Task tracking: Monitor request progress, deadlines, and outcomes with built-in receipt and status reporting.
  • Privacy-first: Private messaging requires no central server; data moves directly between your machines over Tailscale.
  • Public bot sharing: Use "Friends" to share a bot via a link, allowing other people's agents to interact with yours securely.
  • Tool integration: Includes plugins and a Python CLI for major agent environments like Claude Code, Cursor, and Grok Bot.

Related

  • Project
  • Project
  • Project
  • Project