Skip to main content

Overview

Type: IDE and CLI coding agentPrimary Path: OpenAI-compatible Chat CompletionsSupport Confidence: Official custom-provider path
Kilo Code supports custom OpenAI-compatible providers in its VS Code extension and Kilo CLI. Both clients share the same provider and model configuration.

Install

For VS Code, search for Kilo Code in Extensions and install the pre-release channel. Kilo documents that channel as the current stable, recommended extension. For the CLI, run:

Configure in the UI

  1. Open Kilo Code Settings → Providers.
  2. Scroll down and click Custom provider.
  3. Use these values:
Kilo automatically queries AI Sonar’s /v1/models endpoint after the Base URL and key are entered. Click Submit, then choose aisonar/kimi-k3 in the model picker.

Configure the CLI

Store secrets in an environment variable:
Add this to the trusted global config at ~/.config/kilo/kilo.jsonc:
Kilo resolves {env:...} only in trusted global or managed configuration. Do not put an API key directly in a project kilo.jsonc that may be committed to Git.

Verify

Restart Kilo after editing global configuration, then run:
Ask Kilo to inspect a file before approving write or shell actions.

Troubleshooting

Confirm the Base URL ends in /v1, the key is active, and kilo models lists the configured provider. Add kimi-k3 manually if model discovery is unavailable.
Custom models need explicit limit.context and limit.output values. Keep the limits aligned with AI Sonar’s current model metadata.
Confirm tool_call is true, then review Kilo’s permission settings for the current project.