All Gradio Spaces throwing 500 Internal Server Error (UI not loading)

Spaces
1

Hi everyone,

I am unable to access or load the UI on any Gradio Spaces. Whenever I open a Gradio Space, it immediately throws a 500 Internal Server Error instead of loading the interface.

  • Issue: Gradio Space UI does not load at all.

  • Error: 500 Internal Server Error.

  • Scope: Happening across multiple / all Gradio Spaces (not limited to a single repo).

Is anyone else experiencing this right now, or is there an ongoing outage/incident affecting the Gradio hosting infrastructure?

Any updates or fixes would be appreciated. Thanks!

2

Hi @Aryan-x we can take a closer look! Can you please share a link to an example Space?

3

I’ve been getting 500 or 503 intermittently when any of my spaces try to load for 4 hours now. Sample URL’S:

4

The two apps I have on Spaces seem to be down as well. Here is a link to one of them. They are not Gradio

5

Same issue here, since ~02:00 UTC on Sep 29 (still intermittent a few minutes ago):

  • One of my Spaces (Gradio 6.28, zero-a10g hardware) stays in “Running” with completely clean runtime logs, but requests to its .hf.space URL fail intermittently — roughly 10–50% depending on the minute. The failures return your HTML error page (HTTP 502), both for static assets (/assets/.js|css) and for /gradio_api/* endpoints (e.g. /gradio_api/info; users also hit it on queue/join and uploads). The browser surfaces it as: “Could not parse server response: SyntaxError: Unexpected token ‘<’”.
  • I reproduced the same intermittent 502s fetching a static JS asset from a different, public Space (hugging-apps/qwen-image-2-1), so it doesn’t look Space-specific.
  • One example request id from a 502: x-request-id BukWd_. Successful responses include x-proxied-host / x-proxied-replica headers; the 502s don’t, which suggests the proxy never reaches the app.
  • Client-side retries on failed asset/API requests mask most of it, but the failures are clearly upstream.

Happy to provide more details, timestamps, or raw request logs if useful.

6

I can’t share, because it’s not a permanent position it’s (500) a temporary but serious internal issue with Hf spaces (mostly on gradio spaces) and my microservices are also affected by this issue.

Thanks

7

i’ve seen the same 500 error on a few spaces today maybe it’s a temporary outage on the hosting side

8

A 500 Internal Server Error across multiple Spaces can point to a broader hosting issue, and the Beat The Heat website also makes it easy to explore services and Texas service areas or request a free quote.