Launched this week
Peras
Pause Claude Code and Codex before they burn your week
69 followers
Pause Claude Code and Codex before they burn your week
69 followers
Peras puts a line in your Claude Code and Codex week. Set it (peras set weekly 85 pause) and once usage crosses it, the agent's next tool call is refused with a plain reason, so one long task can't burn the rest of your week. Resume whenever you want, from the terminal or the menu bar. Opt-in: your agents can be told their own usage. The CLI and hooks are free and MIT. The native Mac menu bar app is €4.99 once.
Payment Required
Launch Team / Built With
MotifCalm, tough or hype: an AI coach for your wearable data
Promoted
where's the weekly % coming from, does anthropic expose that number somewhere now
Congrats on Peras, Kristiyan. A spend line that stops the run itself, instead of reporting it after, is the part that stayed with me.
@w_fan Thanks Fan, that distinction is exactly what I wanted Peras to solve. Have you had a chance to try it yet? Would genuinely love to hear how it feels in a real workflow.
the PreToolUse hook refusing with a reason the agent actually reads is clever, way cleaner than me checking /usage every hour
i'd probably put teh line at 80 on weekdays. can you set a different line for the weekend??? or one per project???
@thamibenjelloun Thanks! That was the idea: the agent gets told why it stopped, so you don't have to keep checking. One per project works today. In that repo, run peras set --here weekly 80 pause. The line covers that folder and everything under it, and replaces your global pause line there. A different line for weekends isn't in yet. A line can be limited to hours of the day (peras set 09:00-18:00 weekly 80 pause) but not to certain days. Until then, you can save two profiles (peras profile save weekday, peras profile save weekend) and switch on Friday with peras profile use weekend. Lines for chosen days are a good idea and I'll look at adding them.
This hits a real pain point. I've watched an agent loop happily through a whole session's budget on one stubborn test. How do you decide when to pause? Is it a raw token or spend threshold, or do you detect unproductive loops, like repeated edits to the same file? Love the focus on just one sharp utility.
@anirudh_madhavan Thanks man, appreciate it! That exact thing is why I built it, I kept finding my week gone because of one stubborn test.
It's a threshold, but not on tokens or dollars. Peras reads the same usage % that Claude Code and Codex show you, for the 5 hour window and the week. Before every tool call the hook checks your lines, and if one says pause, the call gets refused with the reason and the turn ends.
For your loop case I'd use the per-prompt line: peras set prompt 3 pause. One prompt can't burn more than 3% of your week, no matter what the agent is doing in there. Next prompt starts fresh. There's also a pace line that pauses when you're on track to run out before the reset, and by default it pauses at 90% weekly.
No loop detection though. Peras never reads your prompts or commands, it only watches how fast usage goes up. And honestly 5 edits to the same file is sometimes a loop and sometimes just TDD, so I'd rather cap the spend than guess.