AI coding got dramatically better, but the session stayed single-player.
My cofounder Hudson and I already had our Macs connected with Tailscale. The missing layer was not another IDE or chat app—it was a way to enter the exact Codex or Claude Code conversation already in motion, understand the context, and add a turn with a real human name attached.
That is what mpai does. I sent an attributed prompt from my Mac into Hudson’s existing Claude Code session; Claude received the teammate label and answered inside that exact session.
This is an honest public alpha: macOS + Tailscale, explicit session sharing, Claude Code support, and managed-mode Codex prompting. Standalone Codex stays view-only by default because we will not race an active desktop task for a prettier demo.
It is MIT licensed and built in public. I would especially value feedback on install friction, the sharing model, and the terminal interaction.
PH 用户
That's pretty cool, but what are the actual use cases? We use Claude Code, and on our team, up to 10 people can work on the same project simultaneously, but each person is responsible for their own part. Or is this mainly designed to work around usage limits?
PH 用户
Love the idea of collaborating inside the same AI coding session instead of constantly sharing context back and forth. Feels like a much more natural way to work as a team. Congrats on the launch!
PH 用户
What happens if the host and guest send a prompt at the same time? Is there a turn lock, a queue, or does one just get clobbered mid-stream?
PH 用户
the decision to keep host execution authority is the right call, the second you let the guest drive, you've built a pair programming tool, not a session amplifier. curious how attribution surfaces when one person's turn kicks off a 20-step agent chain: does each sub-action stay attributed to the human who prompted it, or does the chain appear as agent output with a single human-turn anchor?
PH 用户
发布日更新：v0.4.13 已上线。主机端仍然只需安装一次；队友现在只需粘贴一行固定版本的 npx 命令，就能进入完全相同的共享房间——无需全局安装，无需设置，也没有后台服务。两位创始人的 Mac 都已升级，干净的 Homebrew 生命周期在 macOS 和 Linux 上通过了测试。接下来关键指标是外部使用：前 10 个双人团队，邀请到进入房间的中位数时间低于 5 分钟，然后在当周晚些时候再回看数据。测试清单：https://github.com/godfaddaai/multiplayer-ai/issues/7
My cofounder Hudson and I already had our Macs connected with Tailscale. The missing layer was not another IDE or chat app—it was a way to enter the exact Codex or Claude Code conversation already in motion, understand the context, and add a turn with a real human name attached.
That is what mpai does. I sent an attributed prompt from my Mac into Hudson’s existing Claude Code session; Claude received the teammate label and answered inside that exact session.
This is an honest public alpha: macOS + Tailscale, explicit session sharing, Claude Code support, and managed-mode Codex prompting. Standalone Codex stays view-only by default because we will not race an active desktop task for a prettier demo.
It is MIT licensed and built in public. I would especially value feedback on install friction, the sharing model, and the terminal interaction.