Is dots broken for others too or a small percentage have issue using it?
Users are reporting issues with Dots' cloud computer, specifically encountering 403 Forbidden errors when trying to access various websites. This limitation prevents tasks like logging into Cloudflare to set up email routes. There's a concern that Dots' context memory has become fragmented, leading to a perceived decline in its proactive capabilities, reminiscent of bugs observed during its dev day announcement.
Time & source
Times shown in UTC
Display time zone: UTC
Local time zone unavailable; showing UTC.
PublishedOffset at this time: UTC+0Oct 10, 2026, 09:46 UTC
IngestedOffset at this time: UTC+0Oct 10, 2026, 11:00 UTC
- Published
- Oct 10, 2026, 09:46
- Ingested
- Oct 10, 2026, 11:00
- Source type
- Dev community
- Tier
- Community
- Source status
- Healthy
Tier is a per-source editorial setting, not a per-item score.
I’m genuinely asking: browser use isn’t working well on Dots’ cloud computer. Every website I’m interested in using returns a 403 (Forbidden) error when accessed from Dots’ cloud. For example, I couldn’t log in to Cloudflare to set up an email route to my home computer, where I could host a mail server and essentially replicate Grokbot’s email feature, which just shows how limitted the cloud computer is (to be fair everything was blocked on Grokbot's cloud computer when I last used it, but there I could ask it to setup a vpn to my own pc).
When using my connected Windows PC, it feels handicapped when trying to do anything. Browser access crashes, as apparently reported here. Many actions on my PC also require my input for example, every script seems to require elevated access, which shouldn’t be necessary for something as simple as dir.
So, basically, what is the reason for using Dots? I could open a separate agent task that isn’t as buggy and has better access. Or are my expectations for what it can handle simply too high?
Above all, the initial hype around it doing things was genuine. Its context was highly limited to the single workflow we had started prompting it with, and everyone kept emphasizing how “proactive” it was. Now, after a couple of weeks, its context memory has inevitably become fragmented, and it just feels lazy. (Thinking twice it is the same bug they had on dev day announcement. and it is very likely they forgot to restart its context and later on their separate recording when it succeeded it was a fresh dots)