Claude Code Bug triggers Rate limits without usage
Starting an hour ago, i received the following message "API Error: Rate limit reached" in claude code on a 5x Max subscription.
I had not used the model extensively, but accepted it. I waited 10min and asked again on how to go about a localization task on a website. Nothing code intensive, just a Pointer on what path to take given the infrastructure. However the same error message. I checked claude status, i checked HN and started the support bot, i reviewed the API ratelimits. But all seemed normal. And nowhere did it seem like i exceeded. I waited another 30min and tried again. The error message persists, according tot the doc, it should tell me how long to wait. it doesnt.
Anyone else experiencing this?
Based in Switzerland, Europe, on Linux I have seen similar rate limit errors even when usage was low, and in my case it was tied to background sessions and cached conversations. Starting a completely new session or logging out and back in sometimes cleared it. Might be worth trying if you have not already. yeah, starting a new conversation and asking it to review the logs of the old one for context is my current workaround. still its a weird bug One workaround if rate limits keep biting: use a proxy that can failover to alternative providers when one hits limits. I built NadirClaw for this - it classifies prompts and routes to the cheapest model that can handle the task, but also switches providers when one is unavailable. Doesn't help if you specifically need Claude's quality for every task, but for the "pointer on localization" type questions it could route those to a cheaper model entirely and save your Claude quota for the hard stuff. https://github.com/doramirdor/NadirClaw (disclosure: I'm the author) Claude chat has also been buggy the past few weeks in random ways. That they’re heavily vibe-coding their own product is starting to show. Im seeing similar 'You have used 90%' and 'No more free messages' after 1 to 5 prompts on the free web plan during about that same timeframe. It did give a timeout duration of several hours and then reoccured after the time had passed.
That it also happens to a paid plan is not encouraging.