How to set up Kilo Code CLI to use Ollama local? Only lets me select Ollama Cloud

[cleong110]

Question

https://kilo.ai/docs/ai-providers/ollama discusses how to set up Kilo Code in VS Code to use local models via Ollama. However, I am not sure how to get it to work with Kilo Code CLI. It only allows me to select "Ollama Cloud":

I can confirm that ollama is installed and works. curl http://localhost:11434/api/version gives me {"version":"0.17.7"}(base), and ollama list works too:

ollama list
NAME                       ID              SIZE      MODIFIED
glm-4.7-flash:latest       d1a8a26252f1    19 GB     19 minutes ago
lfm2:latest                d6c816d74887    14 GB     18 hours ago
qwen3-coder:30b            06c1097efce0    18 GB     18 hours ago
gemma3:latest              a2af6cc3eb7f    3.3 GB    19 hours ago
nomic-embed-text:latest    0a109f422b47    274 MB    3 months ago
qwen2.5-coder:1.5b-base    02e0f2817a89    986 MB    3 months ago
llama3.1:8b                46e0c10c039e    4.9 GB    3 months ago

Versions:

kilo -v
7.0.41
ollama -v
ollama version is 0.17.7

[kilo-code-bot]

The user is asking how to configure Kilo Code CLI to use a locally-running Ollama instance instead of 'Ollama Cloud'. They have Ollama installed and working locally but cannot find the option in the CLI.

This response was generated by Kilo Auto-Triage and may not be fully accurate. Please verify the information independently.

[vkduy]

You can check this document: https://opencode.ai/docs/providers/#ollama

Kilo code was built on top OpenCode

[glaucopater]

@cleong110 you can set it up adding a custom provider pointing to your local Ollama (default http://localhost:11434/v1/),
When you will be able to see your local model then you need to set the default model in Models setting.
Kilo chat will be showing your model.
Previously the configuration was a bit easier for custom local model indeed.

[ljubomirb]

But why need for "custom"? I think ollama is used pretty much, and I thought it would be reasonable to expect it is as simple as to enter ip:port, and later to be able to choose model at any time like on previous version from some combo like box.

So, can local ollama also be included, please?

[dtaddis]

I would like local ollama added too! But for now, @glaucopater's suggestion works.

[stiv-yakovenko]

I don't see anything relevant here:

[glaucopater]

You need to add a provider after you setup Ollama locally, then you can set the default model (in this example Gemma4) and then finally you will be able to see the custom model in Kilo chat

[hyh19962008]

I think they drop the ollama intentionally to force you to use cloud service.

[DataDrivenAngel]

+1 would be nice if this was easier and/or better documented. Local LLMs aren't the move for serious coding, but would be nice to be able to have this as an option for airplanes. (I wasn't able to get the CLI working with my local ollama setup).