cannot assign model to autocomplete

[mark-hahn]

I have codestral assigned to chat and it works. When I click on Setup Autocomplete Model it just sends me to a page about autocomplete, like what model to choose, but there is nothing saying how to actually assign and use autocomplete. So I'm stuck.

[bdougie]

@mark-hahn sorry to hear you are stuck . Can you share if you are using the Local Assistant experience and not a hosted experience on the Hub. I'd be happy to update the docs after I gather more context.

But assuming you are using the local experience via the config.yaml with ollama. You can add something like this to the yaml.

models: 
  - name: Autodetect
    provider: ollama
    model: AUTODETECT
    roles:
      - chat
      - edit
      - apply
      - rerank
      - autocomplete

The docs have the example with Codestral in the docs https://docs.continue.dev/reference#models

Int he VS code extension, locally you would need to point to a provider. https://docs.continue.dev/customize/model-providers/overview

I also realize a lot of this context could be centralized to provided as a better experience, so any feedback here would be helpful.

[ggrensteiner]

Also completely stuck because the docs are, at best, thin. Adding the model provider with roles does not clear the completely unhelpful Setup Autocomplete model link under Autocomplete, which leads to something without any documentation on use/setup at all.

[mark-hahn]

I'm sorry I gave an incomplete problem report. I want to use codestra cloud with autocomplete in vscode. I'm confused. I've enclosed what I see in the UI. It says "Set up autocomplete model" but when I click on it there is no indication of how to set up an autocomplete model. Can you step me through setting it up? Also I'm not sure if I'm connected to my Mystral account so also step me through checking out my account connection. [image: continue.gif]

…

On Tue, Sep 16, 2025 at 12:25 PM Brian Douglas ***@***.***> wrote: @mark-hahn <https://github.com/mark-hahn> sorry to hear you are stuck . Can you share if you are using the Local Assistant experience and not a hosted experience on the Hub. I'd be happy to update the docs after I gather more context. But assuming you are using the local experience via the config.yaml with ollama. You can add something like this to the yaml. models: - name: Autodetect provider: ollama model: AUTODETECT roles: - chat - edit - apply - rerank - autocomplete The docs have the example with Codestral in the docs https://docs.continue.dev/reference#models Int he VS code extension, locally you would need to point to a provider. https://docs.continue.dev/customize/model-providers/overview I also realize a lot of this context could be centralized to provided as a better experience, so any feedback here would be helpful. — Reply to this email directly, view it on GitHub <#7732 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAGGDP56KVR3TRLYAQJOIK33TBP3NAVCNFSM6AAAAACGNYQNICVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTINBSGA2TKMA> . You are receiving this because you were mentioned.Message ID: ***@***.***>