One queue for every agent you run.
Namu doesn't run your agents. It's the layer above them, pulling every interrupt from Claude Code, Cursor, Copilot and Codex into a single priority decision queue.
Namu doesn't run your agents. It's the layer above them, pulling every interrupt from Claude Code, Cursor, Copilot and Codex into a single priority decision queue.
Claude Code finishes a migration and waits for a yes. Cursor flags a failing test in another repo. Copilot has a draft ready. Codex hits a rate limit. Each one is confident, capable, and completely blind to the other three.
None of them know you're running the rest, so you become the router. You hold the whole picture in your head and re-explain the same rule to a different tool every time you switch windows.
As the agents get more capable, the constraint stops being their output. It becomes your attention. Namu is the layer that manages that, not another agent competing for it.
They're great individually, but I end up being the human middleware, copy-pasting between Claude windows and manually coordinating who works on what.
Namu reads each agent platform's decision and status events and puts them on one board. It does no work of its own, and it never touches your source.
Every interrupt from every running agent lands in one place: the blocker in plain language, the session it came from, and how urgent it is. You decide once, in order, instead of alt-tabbing to find who is stuck.
See what is running, paused, or blocked across every platform at once. This is the one screen a single-vendor tool structurally cannot show, because it only ever sees its own agents. Namu sees all of them because it sits above them.
Set a rule once and Namu holds every agent to it, no matter which vendor the decision comes from. The boundary stops evaporating every time you switch windows.
A CLAUDE.md only governs Claude Code. Switch tools and that context is gone, so you re-explain the same boundary to the next agent by hand.
Standing rules live in Namu instead, above the agents, so one rule applies across every vendor you run. It holds whether Claude Code, Cursor, or Codex is the one that hits it.
Written in plain language. No config file per tool.
Persist cross-vendor by default.
Namu is pre-launch. These are the rates we're opening early access at, sourced from our go-to-market plan. Not live billing yet.
Join the waitlist and run every agent through one queue. One email when your slot opens.