Claude AI Not Working Today? Here's What to Check
If Claude AI is not working today, the cause is almost always one of five things: a service outage on Anthropic's side, you've hit a usage limit, a browser or app-level glitch, a network/regional block, or an account issue like billing or verification. This guide walks through how to tell which one you're dealing with and what to actually do about it, in order of likelihood.
Before you assume the worst, check the basics: reload the page in a private/incognito window, try a different browser, and check whether the problem happens on claude.ai, the mobile app, or both. That single test already rules out half the possible causes.
Step 1: Check if it's a real outage
The fastest way to confirm a widespread issue is to check Anthropic's status page directly, plus a quick search on social platforms for "Claude down" — if a lot of people are reporting the same thing at the same time, it's an outage, not something specific to your account.
Signs it's a genuine outage rather than a local problem:
- Error messages mention server errors (5xx) rather than authentication or rate-limit messages
- The issue is happening across multiple devices and networks you control
- It started suddenly and affects a specific feature (e.g., only image uploads, or only long conversations)
- Other people online are reporting the exact same symptoms at the same time
If it's a confirmed outage, there's nothing to fix on your end — you wait, or you fall back to a different tool for anything urgent.
Step 2: Check for rate limits and usage caps
This is the most common reason Claude "stops working" for regular users. Every plan — free, Pro, or Max — has usage limits that reset on a rolling window. If you've been having long, message-heavy conversations, especially with large file uploads or long context, you can hit the ceiling well before you expect to.
Symptoms of a rate limit rather than an outage:
- A specific message telling you you've reached your usage limit, often with a reset time
- The web app loads fine, but sending a new message fails or gets throttled
- It only happens in one particular conversation (long threads consume more of your quota per message)
If this is the issue, the fix is either to wait for the reset window, upgrade your plan, or start a fresh conversation instead of continuing a very long one — long threads reprocess more context on every message, which burns through usage faster than starting new.
If you're hitting this constantly because you're building something on top of Claude — a script, an internal tool, a product feature — a consumer chat plan isn't really built for that kind of repeated, programmatic use. That's a different problem with a different fix, covered below.
Step 3: Rule out browser and app issues
A surprising number of "Claude is broken" reports are actually local rendering or caching problems:
- Clear your browser cache and cookies for claude.ai
- Disable browser extensions one at a time (ad blockers and privacy extensions are common culprits for chat apps that use streaming responses)
- Update the Claude mobile app to the latest version
- Try a wired connection or different Wi-Fi network if you're on a flaky connection — streaming responses are sensitive to connection drops mid-message
If Claude loads but responses cut off partway through or never finish streaming, this is very often a network stability issue rather than anything on Anthropic's end.
Step 4: Check your account and region
Less common, but worth ruling out:
- Billing issues: an expired card or failed payment can silently downgrade or restrict your account
- Regional availability: Claude isn't available in every country, and VPN usage can trigger access restrictions
- New account verification: some accounts require additional verification steps before full access is unlocked
If you see a message about account restrictions rather than a generic error, this is where to look first.
When the real problem is that you're not using it through the right interface
Sometimes "Claude AI is not working" really means "the consumer chat app isn't reliable enough for what I'm trying to do." If you're pinging Claude repeatedly from a script, a workflow automation tool, or an internal app, you're fighting against interface limits and rate caps that were never designed for that kind of traffic. The actual fix isn't waiting out an outage — it's using the API properly, with your own authenticated key, retries, and usage visibility.
That's the gap SubToAPI closes. It turns your existing Claude access into a proper HTTPS API: you get an application key (sub_live_...), streaming support, tool use, and usage metadata in a dashboard, so instead of debugging a browser tab you get predictable, programmatic access.
curl https://api.subtoapi.app/v1/messages \
-H "Authorization: Bearer $SUBTOAPI_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "claude-sonnet-4",
"max_tokens": 1024,
"messages": [
{"role": "user", "content": "Summarize this outage report in two sentences."}
]
}'
Setup takes a few minutes — see the quickstart guide for the full flow, /docs/messages for request formats, and /docs/streaming if your app needs token-by-token output. Plans start at €9/month with a free trial at /signup, and pricing details are on /pricing.
Quick troubleshooting checklist
- Check Anthropic's status page and social media for outage reports
- Look for a rate-limit or usage-cap message in the app
- Try a different browser, clear cache, or update the mobile app
- Check your connection stability, especially for streaming responses
- Verify billing status and regional availability on your account
- If you're using Claude programmatically, consider moving to a proper API setup instead of the chat interface
questions
Is Claude AI down right now, or is it just me? Check Anthropic's status page and search for recent reports on social platforms. If others are reporting the same issue at the same time, it's a broader outage; if it's isolated to your device, it's more likely a browser, network, or account issue.
Why does Claude stop responding in the middle of a long conversation? Long threads consume more of your usage quota per message because Claude reprocesses the full context each time. This can trigger rate limits faster than expected — starting a new conversation for a new topic usually helps.
How do I avoid hitting Claude's limits when building an app or automation? The consumer chat interface isn't designed for repeated programmatic calls. Using the API directly, or a managed layer like SubToAPI with its own application keys and usage tracking, gives you predictable access without fighting chat-app rate limits.