Incident #030 · opened 2026-05-28 · 20:07 UTC
Ollama enabled but routing stays on "Gerenciadas" (Managed/Cloud) — local model not used despite being toggled on
new
close: support-only
Description
User has Ollama toggled ON under LLM Providers, with an arrow annotation pointing to it. However, the Routing section below shows "Gerenciadas" (Managed) selected — meaning all inference is still routed through OpenHuman Cloud despite the local Ollama toggle being active. The local model will not be used until routing is switched to "Use seus próprios modelos" (Use Your Own Models) and Ollama is selected as the provider.
This is the same UX gap seen in Lobster's ticket and Sir_SabreWolf's ticket: enabling a local or custom provider does not automatically update the routing mode. The user must manually change routing — but there is no prompt, tooltip, or visual warning indicating this. The Ollama toggle appearing "active" creates a false impression that local inference is in use.
Fix needed: when a non-managed provider is toggled on, the routing section should either auto-switch or display a clear warning that routing must be changed manually to use it.
Attachments
screenshots · logs · files · 1 item
Notify the user
copy a Discord-ready message, then mark as notified
Interventions
refunds · credits · manual fixes
No interventions logged.
Activity
4 events
Pooja
pushed to GitHub → issue #3093
May 31 · 21:58
Pooja
changed priority high → urgent
May 28 · 20:08
Pooja
attached image.webp
May 28 · 20:07
Pooja
created the incident priority=high
May 28 · 20:07