All your agents in one place
Sessions from Claude Code, Codex and other CLIs, side by side. Organized by project.

All your agents on one screen. You know which one needs you.
Follow Claude Code, Codex and other CLIs in a single window, without switching terminals.
I want to try the alpha versionAnyone running several agent sessions at once wastes time switching terminals to find out who finished, who got stuck and who is waiting for an answer. Nidavellir tells you at a glance.
Sessions from Claude Code, Codex and other CLIs, side by side. Organized by project.
A traffic light shows who is working, who stopped and who is waiting for an answer, and for how long. When it can't tell, it says so.
Send messages, approve permissions and resume sessions. No need to go back to the terminal.
Plan steps, tools in use and background commands. Subagents show up too.
Context window and tokens for each session. How much quota each CLI has left.
One agent hands work to another, even from a different CLI, and is notified when it ends. The quota of each subscription is shared out.
A conversation started in one agent continues in another. Context is not lost.
Git (stage, commit, push), files, GitHub issues and terminals. Plus notes for each session.
Runs locally with the Claude Code and Codex you already have. Your data stays on your machine.
For Windows.
It detects Claude Code and Codex installed on your computer.
Open your sessions and see everything in a single window.
Leave your contact details. We send the alpha license personally, as we open new spots.
During the alpha, nothing.
The alpha is Windows only. Other platforms will come later.
Nidavellir keeps everything locally. The conversation with the model goes through your own CLI, under your own account.
Yes. It uses the CLIs you already have.
No. You don't set up tokens. Nidavellir uses the CLIs already installed and configured on your computer, such as Claude Code and Codex, talking to them through the ACP protocol.