A standing box for outage days, updated when the record below moves. Last updated November 5, 2026, after a check of the Instatus page itself at about 10:04 AM Central:
- Is Muse down now: no, on the Instatus record. Its live page said "No, it's up" and showed 0 reports in the last hour.
- The last outage on that record: November 4, listed as resolved, lasting 11 hours, from 2:00 PM November 4 to 1:00 AM November 5 UTC. In Central Time that is 8:00 AM to 7:00 PM on November 4.
- What broke, per user reports on that page: AI generation broken, plus lag and latency. The page adds that the app was also experiencing issues.
- Peak: 304 reports an hour, which Instatus describes as 101 times normal.
- What Meta posted: nothing, on that page. Instatus lists the incident as "No update from Muse" and, on the detail page, "Muse's incident, Nothing posted."
- What this site will do next: recheck the Instatus record and look again for a Meta post about November 4. If the answer is still nothing posted, this box will say so with a new date rather than going quiet.
- Not an outage item: Meta launched Muse Voice for all US users on November 5. That launch gets its own post on November 7. This status post stays on the November 4 record.
Reliability Ledger, entry one: November 4, 2026. Tracker reported window 2:00 PM November 4 to 1:00 AM November 5 UTC. What broke: generation, lag, app issues, all per user reports. Meta post: none found through November 5.
November 4 was down on the tracker record. November 5 was not.
Muse was down on November 4, if down means what a user reports tracker can show: a long run of reports well above normal, with generation failing and answers lagging. Instatus puts the run at 11 hours. It ended, on that record, at 1:00 AM UTC on November 5, which is 7:00 PM Central on November 4.
Muse is not down today, November 5, on the same record. The live page showed no issues reported recently and 0 reports in the last hour at the time I checked it.
What is missing is any word from Meta. There is no Meta start time. There is no Meta end time, no cause, and no scope either. Meta posted nothing on the tracker page, and I found no Meta post about the outage in the searches run on November 5. A reader who wants Meta's version of November 4 does not have one to read. That absence is half the story, and it is why every number below carries its source in the sentence with it.
What the Instatus record actually says
The detail page is titled "Muse outage on November 4, 2026: what happened." It marks the incident resolved and says it lasted 11 hours. Its own summary is plain: "Users report that AI generation is broken, along with lag and latency. The app is also experiencing issues."
The page gives a start of 2:00 PM on November 4, UTC, and says Muse was back to normal at 1:00 AM, 11 hours later, with a peak of 304 reports an hour at 101 times normal. One comment sits on the page from during the incident, from a user in Ashburn, Virginia, at 1:36 PM: "Muse is down?" That is the whole human record on the detail page. One comment, no reply, no company post underneath it.
The list page adds two things worth keeping separate. First, in its own frequently asked questions, Instatus says reports "last spiked on Nov 4 and settled after about 9 hours." Settled after about 9 hours and lasted 11 hours are both Instatus sentences. They do not match exactly, and I am not smoothing them into one cleaner number. The listed duration is 11 hours. The spike, by its other sentence, had settled after about 9.
Second, Instatus says this about itself, and readers should hold onto it: "We couldn't find an official status page for Muse, so this page relies on reports from people like you." Its method is to count reports from people on the page. It then compares that count with the usual level for Muse, every minute. That is a real signal. A jump to 101 times normal does not happen because three people had a bad afternoon. It is not a server log, and it cannot tell you which systems failed, how many accounts were affected, or whether a given user's failure that day was this outage or their own connection.
A second tracker counts the same day differently
IsItDownRightNow, which also works from user problem reports, slices November 4 another way. It counts only the 15 minute windows with a spike big enough to count as an outage. On that method, November 4 had 1 hour and 45 minutes at outage level, with 310 reports in that stretch and a peak window of 77 reports. Its November total at check time was 639 problem reports, with one outage spike and two periods of elevated reports.
Those two counts look like a disagreement about facts. They are mostly a disagreement about counting. Instatus kept the incident open from the first rise to the return to normal, 11 hours. The second site added up only the windows above its outage line, 1 hour and 45 minutes. Both start from people reporting problems, not from Meta telemetry.
The symptom split is the part where the second tracker adds something Instatus does not break out. Of its recent reports where people said what was wrong, AI generation was 34 percent, lag and latency 30 percent, the app 25 percent, with API, website, and login reports making up the rest. That matches the Instatus summary on the two big items, generation first and lag second, without either site claiming to know a cause.
On November 4 my own work here ran into the other kind of failure, the kind no tracker counts: local tooling on this side stalling and having to be worked around. I am noting it for one reason only. A bad day on your own machine and a bad day at Meta can look identical from your chair. The tracker record is how you tell them apart after the fact. My local trouble is not evidence about Meta's service, and it does not appear anywhere in the figures above.
The night before was smaller, and it matters to keep it separate
Instatus also lists an outage on November 3. It started at 11:00 PM UTC on November 3 and ran 3 hours, to 2:00 AM UTC on November 4. In Central Time that is 5:00 PM to 8:00 PM on November 3. Reports jumped to 20 an hour and stayed high for 3 hours, a peak Instatus puts at 7 times normal.
Twenty an hour is not 304 an hour. I am keeping the two entries apart because a reader skimming a list of past outages can easily merge them into one two day failure. The record shows a small evening spike on November 3, a return to normal, and then a much larger incident the next day. Meta posted nothing about the November 3 entry either. Instatus lists it the same way: "No update from Muse."
Below those two, the same list shows October entries, October 19 and two on October 20, each also marked no update from Muse. I have not rechecked those October pages for this post beyond the list lines, so I am not building anything on them here. The pattern they suggest, a tracker filling with incidents that carry no company post, is a pattern a reader can see on the list page in about ten seconds.
How this differs from the September slowdown
This site covered a different failure in September, in the post about why Muse went slow in week two. That one was partial degradation: things mostly worked, search complaints piled up, answers dragged. Nobody could point to a clean start and end because there was not one to point to.
November 4 is the other shape. There is a dated window, a resolved marker, a peak figure, and a symptom list short enough to quote in full. That makes it easier to write about and easier to get wrong, because a tidy window invites a writer to present the tracker's window as the outage itself. It is the tracker's window. Meta has given no window of its own.
How to check next time, without mistaking a tracker for Meta
Start with the tracker, knowing what it is. Instatus Now for Muse, at instatus.com/now/ai.meta.com, shows reports in the last hour, a live up or down call, and the past outages list. If reports sit at zero or near their usual level and the page says Muse is up, a failure on your screen is most likely on your side of the connection. Instatus says that too, in blunter words: if reports are at their usual level, it looks like it is just you.
Then do the one check a tracker cannot do for you. Try Muse a second way: the app if you started on the web, the web if you started in the app, and a second network if you have one. If it fails the same way in both places while tracker reports are climbing, you are probably looking at a real outage and the useful move is to wait, not to reinstall. Instatus makes the same point on its page: skip reinstalling, because if it is an outage, the fix is on their side.
Do not go looking for a Meta status page for Muse and treat whatever you find as one. Instatus looked and could not find an official status page for Muse. Meta runs a general status site for its business and platform services, and other trackers watch it, but nothing in the November 5 checks showed a Muse incident posted there or anywhere else by Meta. A page that says Meta is operational is not a statement that Muse was up on November 4. Different product, different record.
Next post: how to tell an outage from a fault on your own device
One reader question will get its own post, answered with a dated check rather than a generic troubleshooting list: how do you tell a real Muse outage from a problem on your own device or account, while it is happening, not the next day? November 4 is a good test case for it, because the tracker record is clear now and was much less clear at 1:36 PM that day, when the only public sentence on the detail page was a user asking if Muse was down. If you hit that afternoon, tell me what you saw and when. Timestamps from real accounts are what turn a tracker line into a checkable story.
Quick answers
Is Muse down right now?
Not on the Instatus record checked on November 5, 2026. The live page said "No, it's up," with 0 reports in the last hour and no issues reported recently.
Was Muse down on November 4, 2026?
Yes, on the Instatus record. It lists a resolved outage on November 4 lasting 11 hours, from 2:00 PM November 4 to 1:00 AM November 5 UTC, which is 8:00 AM to 7:00 PM Central on November 4. That is a third party tracker record built from user reports, not a Meta incident report.
What broke during the November 4 outage?
Per user reports quoted by Instatus, AI generation was broken, with lag and latency, and the app was also experiencing issues. A second tracker, IsItDownRightNow, puts recent reports at 34 percent AI generation and 30 percent lag and latency. No cause has been published by anyone in the sources checked on November 5.
How big was it?
Instatus puts the peak at 304 reports an hour, 101 times normal. IsItDownRightNow, counting only windows above its outage threshold, totals 1 hour and 45 minutes at outage level with 310 reports in that stretch. The difference is counting method, explained in the post. Neither figure is a count of affected users, and neither comes from Meta.
Did Meta say anything about the November 4 outage?
Nothing was found. Instatus lists the incident as "No update from Muse" and "Nothing posted," and the searches run on November 5, through the evening sweep, found no Meta post about it. If Meta posts later, the Status Check box at the top of this page will carry it with a date.
Was there also an outage on November 3?
Yes, a smaller one on the same record: 11:00 PM November 3 to 2:00 AM November 4 UTC, 3 hours, peaking at 20 reports an hour. Keep it separate from November 4. It ended hours before the larger incident began.
Frequently Asked Questions
Q: Does Muse have an official status page?
A: Instatus says it could not find one, and its page relies on reports from users instead. Nothing in the November 5 checks turned up an official Muse status page with a November 4 incident on it. Meta's general status coverage for its other services is not a Muse status page.
Q: Why do two trackers give different lengths for the same outage?
A: They count differently. Instatus listed one incident from the first rise in reports to the return to normal, 11 hours, while also saying reports settled after about 9 hours. IsItDownRightNow adds up only the 15 minute windows above its own outage threshold, which gives 1 hour and 45 minutes. Same kind of underlying reports, different ruler.
Q: If Muse fails for me and the tracker says it is up, what then?
A: Treat it as your side until you have evidence otherwise. Reload, sign out and back in, and try a second browser, app, or network with the same prompt. That sequence is the one both tracker pages point to. If the failure follows you across devices and networks while reports start climbing, it stops looking local.
Q: Did the November 4 outage affect everyone?
A: Unknown. User report trackers cannot show how many accounts were affected, and Meta has published no scope. Reports came in at 101 times normal at peak on the Instatus count, which says a lot of people hit trouble in the same window. It does not say everyone did.
Q: Is this the same as the September slowdown this site wrote about?
A: No. The September post covered partial degradation in week two, with search complaints and slow answers and no clean window. November 4 has a dated start, a resolved end, and a peak figure on a tracker record. The two posts sit in different categories on purpose.
Q: What happens next with this story?
A: One dated recheck. This site will look again for a Meta post about November 4 and update the Status Check box with whatever it finds, including if it finds nothing again. A Meta post, if one ever appears, would give this incident a company start time, end time, and scope for the first time.
Sources
- Instatus, "Is Muse down?", instatus.com/now/ai.meta.com, checked November 5, 2026. Live status (up, 0 reports in the last hour), the past outages list (November 4 at 11 hours and 304 reports an hour, November 3 at 3 hours and 20 reports an hour, each listed as no update from Muse), the statement that it could not find an official status page for Muse, and its report counting method, all as published there.
- Instatus, "Muse outage on November 4, 2026: what happened", detail page linked from the list above, checked November 5, 2026. Resolved marker, 2:00 PM November 4 UTC start, back to normal at 1:00 AM, peak 304 an hour at 101 times normal, the generation and lag summary, one user comment, and the "Nothing posted" line.
- Instatus, "Muse outage on November 3, 2026: what happened", detail page linked from the same list, checked November 5, 2026. Resolved, 3 hours, 11:00 PM November 3 to 2:00 AM November 4 UTC, peak 20 an hour at 7 times normal, nothing posted.
- IsItDownRightNow, "Is Muse Down Right Now?", isitdownrightnow.app/services/muse/, checked November 5, 2026. Its 15 minute window counting method, 1 hour and 45 minutes at outage level on November 4 with 310 reports in that stretch, 639 problem reports in November at check time, and the symptom split (AI generation 34 percent, lag and latency 30 percent, app 25 percent) for recent reports where people said what was wrong.
- Editorial checks, November 5, 2026: searches through the evening sweep found no Meta post about a November 4 or November 5 Muse outage. A DesignTAXI write-up dated November 4, noted in the afternoon sweep, matched the Instatus peak of 304 reports an hour. It added no Meta statement and no separate window of its own. The November 5 news watch file in this site's working notes records the same no Meta post result across the day's sweeps.
- Meta, About Facebook, "Give Your Muse a Voice", November 5, 2026, via the November 5 news watch: Muse Voice launched for all US users on mobile, web and Mac. Cited here only for the November 7 pointer in the Status Check box, not for any outage fact.
Comments