What happened
Bolt.new had a bigger problem starting at 06:35 UTC on June 23, 2026. It lasted about 9 hours and was resolved at 16:07 UTC. Bolt.new has since confirmed the issue is resolved. The details of what went wrong have not been published beyond that.
Who this kind of outage hits, and how they usually find out
If you build something on Bolt.new, a 9-hour disruption is a long time for work to quietly stop. Most people who run automations or apps on a tool like this find out the same way: a customer writes in, confused or annoyed, and only then does the operator start digging. By that point the problem has been going on for hours. There was no alert, no red screen, nothing that said “stop, something is broken.” Just silence, and then an unhappy message from someone who was depending on the thing you built.
Why this is especially rough if you are not an engineer
When a tool you build on goes quiet, there is nothing obvious to look at. No log file, no error message sitting on your screen. The work just stops moving. If you are running an app or a set of automated workflows, the first sign is often that something that should have happened did not. A form submission that never triggered a follow-up. A project that never generated. A customer who waited and got nothing. You start second-guessing yourself, wondering if you broke something, before you even think to check whether the tool itself is the problem. That gap, between the tool going down and you finding out, is where the damage happens.
Timeline
- 06:35 UTC, June 23, 2026 - Bolt.new begins reporting a bigger problem on its public status page.
- About 9 hours of disruption - the problem continues through the morning and into the afternoon UTC.
- 16:07 UTC, June 23, 2026 - Bolt.new marks the issue resolved.
How a watcher catches this before your users do
NoCrash reads Bolt.new’s public status page every minute. The moment that page flips from working to having trouble, NoCrash sends you a plain-language message, in words you can act on, without you having to go check anything. On a 9-hour outage like this one, that means you hear about it close to 06:35 UTC, not at hour seven when a customer finally writes in.
NoCrash also watches the things you ship. If you have n8n workflows, it watches those. If you have an app, you can give it a URL or drop in a small JS snippet and it will watch that too. So if something goes quiet on your own side, that surfaces as well. It is all in one place, in plain English, without you needing to watch a dozen different status pages yourself.
What it does not do: it does not find the outage before Bolt.new’s own status page reports it. It reads that page and tells you quickly. That is the honest version of what it is.
The official source
For the authoritative account of this outage, go to Bolt.new’s own status page: https://status.bolt.new/proxy/status.bolt.new