Default config in VSCode cannot be changed - How to add models from hub to VSCode extension

[mljenkinsdotcom]

In VSCode, if I open Continue, click the top-right gear icon to get to Settings, click Configs, and click the Gear icon next to Default Config, it takes me to https://www.continue.dev/continuedev/default-config which does not exist.

When I select Default Config, I can see the models that are setup at https://www.continue.dev/settings/models although a few of them say missing API key when I look at them in the extension settings under models.

On a side note, the local config is currently empty.

Is there a way to edit the default config?

Is there a way to auto import the models that are selected in the hub into the extension config?

I was under the impression when I purchased credits on the hub that it would be easier to Continue with the various models within the Extension. Was I incorrect on this assumption? It seems it may have been easier for me to just setup the model under the local config and create and input the API key for the model I want to use.

I confused about how to use the credits I purchased via the hub with the extension.

PS: On a side note, there doesn't seem to be a way to delete a config in the extension. There is no easy way to delete one you create and don't want anymore.

Thanks for the input.

[rehan243]

So is the confusion mainly around how the extension ties into the hub credits/models setup? Because yeah, the behavior you're describing with the missing API keys and the default config redirect is a bit janky.

For editing the default config, I don't think the current UI supports it directly (and that broken link you mentioned confirms it). You'd probably need to manually edit the config.json file. Check in your local extension storage, usually under something like:

~/.vscode/extensions/<continue_folder>/configs/config.json

You can add models or set API keys there directly. Example format (adjust for your model):

{
  "models": {
    "openai-gpt-3.5": {
      "apiKey": "your-api-key-here",
      "endpoint": "https://api.openai.com/v1/completions"
    }
  }
}

For the hub models auto-importing into the extension config—doesn't seem like that's implemented in the current version. You're right, managing it manually in the local config might be simpler for now. Auto-sync would be a killer feature, though—did you check if there’s an open issue for that? Might be worth suggesting.

Deleting configs is another legit pain point. We hit this kind of UX gap in our own tools before, and honestly, the quickest workaround might be nuking the config file directly. Not ideal, but it works.

Curious, are you on the latest extension version? Sometimes they slip minor fixes into releases without documenting them super clearly.