AI Briefing
KOSign in

Jotbus launches as an encrypted shared scratchpad for coding agents

·2026.10.06 22:45

Key point

The tool enables context sharing between agents like Claude Code and Codex via a simple npx command, with a free 60-minute workspace tier.

Details

Jotbus is a new tool designed to facilitate context sharing between coding agents without requiring complex orchestration frameworks or account signups. By running npx jotbus, users create an end-to-end encrypted workspace that allows agents such as Claude Code, Codex, Cursor, and Gemini CLI to exchange notes, files, and instructions across different machines.

How It Works

The system operates on a simple join-and-prompt model:

  • Setup: Running npx jotbus creates a temporary workspace and automatically configures selected agents.
  • Joining: Other agents or users can join using a generated invite command (npx jotbus join ...) without needing an account.
  • Communication: Agents exchange messages via @mentions. A message sent to an agent is delivered the next time that agent is active, rather than waking it from an idle state.

Security and Privacy

Jotbus emphasizes privacy by ensuring that decryption keys never reach the server. Messages and files are encrypted on the user's local machine before transmission, and the service stores only ciphertext. The platform does not scan repositories, transcripts, or shell history; only content explicitly written to the workspace is uploaded.

Pricing and Availability

  • Instant (Free): Provides a 60-minute workspace with no signup required, supporting files up to 2 MB.
  • Personal ($7/month): Offers persistent workspaces, durable history, and increased file limits (25 MB per file, 250 MB total).

The tool requires Node 20 or later and supports any MCP client, though it specifically integrates with major coding agents like Windsurf, Qwen Code, Amp, Augment, Kiro, and JetBrains Junie.

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.