Skip to main content
With one Capriole AI API key, you can run GPT-6 Astra, Claude Fable 5.1, and Opus 5.5 in Kilo Code at up to 92% less than official API prices.

What is Kilo Code

Kilo Code is an AI coding agent for VS Code and the terminal. This setup adds Capriole AI as one custom provider. Each model in it uses the API protocol that Capriole AI serves for that model family.

Prerequisites

Before you begin, make sure you have:
  • A Capriole AI account (Sign up here)
  • A Premium plan (USD 9 per month, also includes unlimited web chat) or a Team plan, required for an API key (Subscribe)
  • A Capriole AI API key (Get your key)

Installation

Kilo CLI

Verify the install:

VS Code extension

Install Kilo Code from the VS Code Extensions view. The extension reads the same global config file as the CLI.

Configuration

Model validation

Kilo CLI 7.8.15 passes a file-read, file-write, and final-answer check through Capriole AI with every model in Supported models, using the configuration below. Create or update ~/.config/kilo/kilo.jsonc:
What each setting does:
  • npm: "@ai-sdk/openai-compatible" is the default package for the provider, so Gemini, Grok, GLM, and Kimi use Chat Completions.
  • Each GPT model overrides provider.npm with @ai-sdk/openai, which uses the Responses API. Each Claude model overrides it with @ai-sdk/anthropic, which uses Anthropic Messages.
  • {env:CAPRIOLE_AI_API_KEY} reads your key from the environment. The Authorization header sends it as a Bearer token for every package. Without it, the Anthropic package sends the key in an x-api-key header, which Capriole AI does not accept.
  • tool_call: true lets the model call Kilo’s tools.
  • limit sets each model’s context window and output limit. Without it, Kilo treats both as 0 and does not compact long sessions.
Keep this provider in your global config. Kilo does not resolve {env:...} references in a project-level config. Run kilo config check after you edit the file.

Run Kilo Code

Set your API key, then start Kilo from your project directory:
Start with a specific model:
For a one-off request, pass a prompt to kilo run:

Supported models

Model references use the form capriole-ai/<model-id>. openai-latest and claude-latest always point to our latest flagship in each family, currently GPT-6 Astra and Claude Fable 5.1.

Troubleshooting

Why does Kilo Code return an authentication error?

Confirm that CAPRIOLE_AI_API_KEY is exported in the same shell before you run kilo, and that the provider is in ~/.config/kilo/kilo.jsonc, not in a project config.

Why do Claude requests fail with Unsupported Claude Messages model?

Claude models that use @ai-sdk/anthropic need the bare Claude IDs, such as claude-opus-5-5. The anthropic/-prefixed IDs are for Chat Completions.

Why does Kilo Code say the model is missing?

Confirm that the selected model uses the capriole-ai/<model-id> format and that provider.capriole-ai.models has a matching entry. Run kilo models capriole-ai to list the models Kilo loaded.

Why does tool use fail?

Confirm that the model entry includes "tool_call": true.

Notes

Last modified on November 9, 2026