'Index not loaded'

[chriswhite5918]

• Disable all other plugins besides Copilot (required)
• Screenshot of note + Copilot chat pane + dev console added (required)

Copilot version: 2.7.2

Describe how to reproduce

I’m using the desktop version. After installing the plugin, configuring the API, and restarting the plugin as instructed, I found that QA mode isn’t working, with the message 'Index not loaded.' Restarting the plugin didn’t help either. I tried forcing the index using a shortcut, but it showed the message 'Indexing is disabled on mobile devices.'

I then created a new empty vault and only installed your plugin, but the result was the same.

Expected behavior

The expected outcome should be that indexing works normally. Even if there’s an issue with the API, it should indicate an API-related problem rather than saying 'Indexing is disabled on mobile devices.

Screenshots

[logancyang]

#825 was merged. Should be fixed in the next release v2.7.3.

In the meantime the solution is to refresh with the dev console open.

[chriswhite5918]

#825 was merged. Should be fixed in the next release v2.7.3.

In the meantime the solution is to refresh with the dev console open.
I updated to version 2.7.3, but the same issue persists. Then I tried shortcuts like Ctrl+R and Ctrl+Shift+R to refresh with dev console open, but the issue still remains after refreshing.

[logancyang]

@chriswhite5918 Could you turn on debug mode as the issue template suggests? The debug messages might help. For now I can't reproduce this behavior on my end.

What's your OS and what embedding model do you have?

[chriswhite5918]

@chriswhite5918 Could you turn on debug mode as the issue template suggests? The debug messages might help. For now I can't reproduce this behavior on my end.

What's your OS and what embedding model do you have?

OS: Windows 11 Pro 26100.2314
embedding: text-embedding-3-small
Looks like the latest version fixed the 'Index not loaded' bug. It seems my model is having issues with this plugin, though it's working fine on Dify. I might need to spend some time figuring out what's going on.

[logancyang]

@chriswhite5918 what model are you using?

[onaimad]

I have the same issue, even on latest version, when using a local model.
It seems API key is mandatory when using 3rd party provider of local model in QA mode. The exact same model works perfectly when added (w/out API key) in chat mode.
OS: Mac Os 15.1
local model: llama3.2-3b-instruct
url: http://localhost:1337 (Jan.ai)

Monitoring the log of the local model server, there are no requests being made if API key is not provided (in QA mode), whereas if using a random key (any string given it's not necessary) the request is sent correctly to the server but it obviously get rejected as the API key parameter is not needed.