ChatGPT Troubleshooting Checklist: Fixes for Request Failures, Endless Loading Spinner, and Unable to Send Messages

If ChatGPT won’t open, keeps spinning, or won’t send messages, it’s usually not because the “model is broken,” but because restrictions are being triggered by your network, browser environment, or account status. Below is a ChatGPT troubleshooting process ordered from fastest to slowest—this should generally help you pinpoint the cause and provide actionable fixes. First, do 3 quick checks (time-saving ChatGPT troubleshooting) Step 1: Check OpenAI Status to confirm whether there’s a server-side disruption; if the status is abnormal, you can only wait for recovery. Step 2: Switch to another network (e.g., a mobile hotspot) and try again—this quickly tells you whether your local network or proxy routing is the cause. Step 3: Log in once using an incognito/private window. If it works in incognito, the issue is usually with cache, cookies, or extensions—this type of ChatGPT troubleshooting is the easiest to fix. Browser and extension conflicts: won’t load, white screen, buttons don’t work When doing ChatGPT troubleshooting, first disable extensions such as ad blockers, script managers, privacy/anti-tracking tools, and auto-translation, then refresh and test. Next, clear site data (cookies + cache) and log in again—many “endless loading spinners” and “blank pages” will disappear immediately. If you’re on a corporate or campus network, check whether a gateway is blocking WebSocket/HTTPS. Switching browsers (Chrome/Edge) can often bypass compatibility issues as well. Network and proxy routing: request errors, unable to send, repeated disconnects Many “Something went wrong” errors are actually caused by an unstable connection. ChatGPT troubleshooting should focus on whether your proxy routes chat.openai.com / chatgpt.com through different nodes. It’s recommended to enable “global mode” or ensure the same domain always uses the same egress, and avoid frequently switching nodes, which can invalidate sessions. If CAPTCHAs keep popping up or login keeps redirecting back and forth, it’s usually due to frequent IP changes or the browser blocking third-party cookies—this points you back to the environment-related ChatGPT troubleshooting in the previous section. Account and usage limits: 429 rate limiting, 403 forbidden, conversations suddenly cut off When you see 429, it’s usually because you’re sending requests too frequently or triggering rate limits in a short period. The ChatGPT troubleshooting approach is to slow down your sending frequency, reduce the number of concurrent tabs, and wait for a while before trying again. 403 is often related to your network egress, region/risk-control policies, or unusual login behavior—first confirm the account can log in stably on the same network, then avoid frequently switching devices and IPs. If a response cuts off or stops midway, split long content into smaller chunks and reduce how much you paste at once—this is an input-side ChatGPT troubleshooting technique. Still not resolved: submit clues using a “minimal reproducible” approach If you’ve completed the above ChatGPT troubleshooting steps and it still fails, record: the exact error message, the time it occurred, your browser version, the network type (direct/proxy), and whether it can be reproduced in incognito. Submitting all of this to official support or a feedback channel is far more likely to be diagnosed than simply saying “it doesn’t work.” Most issues ultimately fall into three categories: server-side disruptions, environment conflicts, or account/risk-control restrictions—troubleshooting against these will converge faster.