Why Is Anthropic Claude Not Working? Common Fixes

When Claude "isn't working," it's rarely one single problem. It could mean the chat interface won't load, a message fails to send, the API is returning errors, or responses are cut off or nonsensical. Each of these has a different cause and a different fix, and most of them have nothing to do with Anthropic's servers being down.

The short answer: check whether the problem is your account/browser, your network, your usage limits, or the underlying service — in that order. Most "Claude not working" reports turn out to be rate limits, expired sessions, browser extensions, or a malformed API request, not an actual outage.

Step 1: Confirm it's not a real outage

Before troubleshooting anything on your end, rule out a service-wide issue. If Claude is genuinely down, nothing you do locally will fix it.

• Check Anthropic's status page for active incidents.
• Search social media for other users reporting the same issue at the same time.
• Try a different network (mobile hotspot) — if it works there, the problem is local to your original connection.

If there's no active incident and other people aren't complaining, the problem is almost certainly on your side — which is actually good news, because it means you can fix it yourself.

Step 2: Browser and app issues (claude.ai)

If you're using Claude through the web interface or app, most "not working" reports fall into a handful of categories:

• Stuck loading or blank screen — usually a stale cache or a browser extension (ad blockers, privacy tools, VPN extensions) interfering with WebSocket connections. Try an incognito/private window with extensions disabled.
• Messages not sending — often a lost session. Log out and back in rather than just refreshing.
• Responses cut off mid-answer — can be a network interruption during streaming, or you've hit a length/context limit for that conversation.
• "Claude is unable to respond" errors — sometimes caused by content that trips safety filters, sometimes by a temporary backend hiccup. Try rephrasing or starting a new conversation.

If none of that helps, clear cookies for the site specifically, or try a completely different browser to isolate whether it's a local configuration problem.

Step 3: Account and subscription issues

Claude access problems are sometimes just account-state issues that look like technical failures:

• Usage limits reached — free and Pro tiers have message caps that reset on a rolling window. If you're suddenly blocked mid-conversation, this is the most common cause.
• Billing failure — a declined card can silently downgrade or suspend paid features without an obvious error message.
• Region or availability restrictions — some features roll out gradually or aren't available in all countries yet.
• Expired session token — especially if you've been idle for a long time or switched networks (e.g., VPN toggled on/off mid-session).

Logging out completely and back in resolves a surprising number of these.

Step 4: API-specific problems

If you're building on Claude's API rather than using the chat interface, "not working" almost always means an HTTP error code, and the fix depends on which one:

• 401 Unauthorized — invalid or expired API key. Regenerate it.
• 429 Too Many Requests — you've hit a rate limit. Add retry logic with exponential backoff instead of hammering the endpoint.
• 400 Bad Request — malformed request body, usually a missing required field or wrong content type for a message.
• 500 / 529 — server-side error or the model is overloaded. These are transient; retry with backoff.
• Timeouts on streaming responses — often a client-side issue with how your code consumes server-sent events, not the API itself.

A minimal, correctly-formed request looks like this:

curl https://api.anthropic.com/v1/messages \
  -H "x-api-key: $ANTHROPIC_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "claude-opus-4",
    "max_tokens": 1024,
    "messages": [{"role": "user", "content": "Hello, Claude"}]
  }'

If this fails, the response body will tell you exactly what's wrong — read it before assuming it's a broader outage.

Step 5: When the problem is really key and rate-limit management

A lot of teams building product features on top of Claude run into "not working" symptoms that are really operational problems: one shared API key hitting rate limits across multiple users, no visibility into which feature or teammate is burning through quota, or no clean way to issue separate keys per environment.

This is exactly the layer SubToAPI sits on top of your existing Claude access. It turns your account into a proper HTTPS API with individual application keys (sub_live_...), so you can isolate a broken integration to one key instead of debugging your entire setup, see usage per key in a dashboard, and manage streaming, tool use, and team seats without building that plumbing yourself. If you're already paying for Claude and just need a stable API layer with keys you can rotate and monitor independently, it's worth a look — see the quickstart or pricing.

const response = await fetch("https://api.subtoapi.app/v1/messages", {
  method: "POST",
  headers: {
    "Authorization": `Bearer ${process.env.SUBTOAPI_KEY}`,
    "content-type": "application/json"
  },
  body: JSON.stringify({
    model: "claude-opus-4",
    max_tokens: 1024,
    messages: [{ role: "user", content: "Hello, Claude" }]
  })
});

Quick checklist

1. Check the status page for active incidents.
2. Test in an incognito window with extensions off.
3. Log out and back in to refresh your session.
4. Check usage limits and billing status.
5. If using the API, read the actual error code and response body.
6. Isolate the failure to a specific key or integration if you're running multiple apps against Claude.

FAQ

Why does Claude say it's unable to respond to my message? This usually means either your account hit a usage limit, the request tripped a content filter, or there was a temporary backend error. Try rephrasing, starting a new conversation, or checking your usage limits in account settings.

Why do I get a 429 error from the Claude API? A 429 means you've exceeded your rate limit for requests or tokens per minute. Implement retry logic with exponential backoff, and if this happens often, consider splitting traffic across separate API keys so one workload doesn't throttle everything else.

Is Claude not working the same as Claude being down? No. "Down" means the service itself is unavailable for everyone, which you can confirm on the status page. "Not working" is broader and more often caused by browser issues, expired sessions, usage limits, or malformed API requests on your end.