Manus is a general-purpose autonomous AI agent from Butterfly Effect, the Singapore-headquartered startup behind the Monica assistant. It takes on whole multi-step jobs such as research reports, slide decks, websites, data analysis, and web automation, and works through them inside a cloud sandbox that you can watch and take over. Because tasks run for many minutes and consume credits as they go, an outage does not just block you, it can waste work already in progress. Start with the live verdict at the top of this page to see whether the problem is widespread right now.
Common reasons Manus stops working
- High service load blocking new tasks. At peak hours Manus can refuse to start anything new, reporting that tasks cannot be created due to current high service load, even though manus.im itself loads normally.
- Long-running tasks crashing mid-execution. Jobs that run for twenty minutes or more can die silently partway through, especially large text or complex multi-step work, with no automatic recovery.
- Sandbox disconnects. The live view of Manus’s computer can freeze and refuse to reconnect, leaving the agent apparently stuck until the sandbox is force-reset.
- The agent stalling at CAPTCHAs and login walls. Browsing steps pause when a site demands verification or a sign-in, and the task then waits for a manual takeover that is easy to miss.
- Credits running out mid-task. Every agent action consumes credits, a task that exhausts your balance stops where it is, and failed runs still burn the allowance.
- Login and region restrictions. Sign-in failures, plus the “Manus is not available in your region” block that has applied to mainland China since mid-2025, can shut some users out while the service is healthy elsewhere.
How to fix Manus when it is not working
- Check the status at the top of this page first. If our check and the official feed show an incident, the problem is on Manus’s side and waiting is usually the fastest fix.
- Retry task creation after a short wait. High service load lockouts are temporary. Trying again a few minutes later, or outside peak hours, often works.
- Reset a frozen sandbox. If the live view of Manus’s computer is stuck or disconnected, force a sandbox reset rather than waiting for it to recover on its own.
- Check whether the agent is waiting on you. Open the live view and look for a CAPTCHA, paywall, or login prompt. Taking over the browser manually unblocks the run.
- Check your credit balance. A task that stopped mid-run may simply have exhausted your credits. Top up or wait for the daily refresh before rerunning.
- Split large jobs into smaller tasks. Shorter tasks are less likely to crash mid-execution and waste fewer credits when something does go wrong.
- Rule out your own setup. Sign out and back in, try a private window or another network, and confirm your region is supported before assuming an outage.
For a fuller checklist that applies to any AI service, see our guide on what to do when an AI tool is down.
Is Manus down, or is it just you?
Compare your experience with what others are reporting. If the verdict and reports at the top of this page show a spike, the outage is shared and the fix is on Manus’s end. If reports are quiet, the cause is more likely local: an exhausted credit balance, a stalled sandbox, a browsing step waiting for takeover, or a regional block. Starting a brand new task is the fastest single test. If a fresh, simple task runs while the original stays stuck, the problem lives inside that one job.
How we check Manus
We determine Manus’s status from two signals rather than one. We read Manus’s official status page, which publishes machine-readable incident updates, and we also run our own independent reachability check against the service from the outside on a regular interval. Combining both means we catch problems even when one source is slow to update, and it keeps the verdict honest. You can read more about how this works on our methodology page.
Alternatives to Manus during an outage
If Manus is down, general assistants can cover many of the same jobs even without a full autonomous sandbox. ChatGPT and Claude both handle research, writing, coding, and analysis well, and Perplexity is a strong stand-in for the research and web-browsing side of a Manus workload. Confirm your backup is healthy before switching, since AI services sometimes wobble together during broad infrastructure events.
Why Manus can fail while manus.im stays up
Manus is really several moving parts: the manus.im web app, the API at api.manus.im, the cloud sandbox where the agent actually works, mobile apps, Scheduled Tasks, and site hosting on manus.space. These pieces fail independently, so the most common Manus complaint is not a dead website. It is a healthy-looking site that refuses to create tasks under high load, or a sandbox that disconnects while the page around it keeps rendering.
Two other patterns are distinctly Manus. Tasks are sometimes reported as completed when the work was not actually done, or the deliverable arrives broken, such as an empty ZIP export. And because every browsing step, code run, and file operation consumes credits, interrupted or failed runs still drain your balance with nothing to show for it. Mainland China is a separate case, since Manus has blocked access there since mid-2025 and shows a region message instead of the app.
- Task creation rejected for high service load while the site loads fine
- Live sandbox view frozen or disconnected until it is reset
- Task marked complete but the deliverable is missing or broken
- Credits drained by a run that failed or was interrupted