To solve the inconvenience of not being able to check AI model usage limits (quotas), developers are creating their own tracking tools to manage consumption.
Imagine this: you’re working hard on a crucial coding project this morning, firing off questions to an AI. Suddenly, the AI sends a cold message: “Sorry, I can’t talk anymore.” You thought you had plenty left, but you burned through your entire quota in just 10 minutes. Why did this happen? How much did you actually use?
Recently, a developer’s story about being unable to stand this frustration and building their own solution went viral on Hacker News: Show HN: My Claude quota ran out in 10 minutes, so I made a tool to find out why
Why does this matter?
AI has become an essential assistant in our daily lives. But just as AI services aren’t free, there is a clear “limit” to how much we can use in a day. The problem is that it is very difficult for us to accurately grasp this limit on our own.
Users often use AI without knowing how much they’ve consumed or when they’ll have full capacity again, only to be caught off guard when the service stops mid-task. It’s like driving on a highway without any idea how much fuel is left in your car. In an era where AI-driven productivity is more important than ever, such opaque usage environments are a major bottleneck that abruptly disrupts a user’s workflow.
Easy to understand: Rotating sushi and entry tickets
Why does this happen? Simply put, AI services manage our usage by giving us “entry tickets” for each day or time period.
Services like Claude Code operate on a “5-hour rolling token usage window.” Claude Code Tool - Check how much of your quota is wasted (DracoMeter) - I made this If we compare this system to a rotating sushi restaurant: if you are using AI right now, the total number of tokens (the word units the AI perceives) you have used in the “last 5 hours” must not exceed a certain threshold. As time passes, the tokens you used earliest drop off the sushi conveyor belt, freeing up capacity again.
However, there is a very important trap here. If you upload the same files to multiple chat windows and ask questions, the AI recognizes these files as new each time and deducts tokens again. How I Stopped Hitting Claude’s Usage Limits In other words, even if you are referencing the same document, the AI calculates it as if it’s reading a new book from page one every single time. It’s like wasting “energy (tokens)” to read the same book from page 1 to 100 every time you need to find a specific piece of information.
In the end, we are burning through our precious “entry tickets” very quickly without even realizing it.
Current situation
Currently, major AI platforms are very closed about users’ token consumption history. Anthropic (the maker of Claude) does not provide detailed analysis data on how many tokens a user has consumed or which conversations consumed the most. Claude Code Rate Limits & Usage Quotas Explained (2026)
That is why people who feel frustrated, like the developer in this case, are building their own “usage tracking tools.” Tracking Claude, Codex, and Gemini Quotas from One Script They write their own scripts to log their AI usage in JSON files or visualize how much they are wasting, gradually correcting their AI usage habits.
Of course, the “Please try again soon” message we see occasionally does not necessarily mean a service failure. It’s just the system temporarily putting you on hold to manage overall user demand; it’s not that the system itself is broken. Troubleshoot Claude error messages However, in these situations, users have no choice but to feel frustrated and yearn for more transparent information.
What’s next?
The AI usage environment is expected to become more transparent in the future. As user demand grows stronger, AI services are likely to provide their own usage management tools or update features to help developers optimize their own usage.
What is the best thing we can do right now? First, actively utilize the “Projects” feature to upload files once and share them across multiple chat windows. How I Stopped Hitting Claude’s Usage Limits Also, it is wise to identify other AI tools in advance in case of AI usage restrictions, or consider subscription-based APIs. Claudeusage limit reached: The Complete Guide for…
MindTickleBytes AI Reporter’s View
As much as AI is getting smarter, managing how “well” we use that AI has become very important. Until the day platforms show usage more transparently, I believe the process of us becoming smart AI users and utilizing tools to manage our consumption is a necessary change.
References
- Tracking Claude, Codex, and Gemini Quotas from One Script
- Claude Code Tool - Check how much of your quota is wasted (DracoMeter) - I made this
- Troubleshoot Claude error messages
- How I Stopped Hitting Claude’s Usage Limits
- Claude Code Rate Limits & Usage Quotas Explained (2026)
- Show HN: My Claude quota ran out in 10 minutes, so I made a tool to find out why
- Claudeusage limit reached: The Complete Guide for…
- Resets at midnight every day
- 5-hour rolling window
- Fixed monthly token amount
- Tokens deducted only once
- Tokens deducted every time it is uploaded
- Unlimited use regardless of file size
- System server failure
- User account suspension
- Temporary limitation due to increased overall user demand