Configuring OpenCode

OpenCode is an open-source terminal-UI (TUI) coding agent that works with any OpenAI-compatible model backend. The recommended way to connect wylon is via a custom provider.

1. Install / upgrade

# Install (if not already installed)
curl -fsSL https://opencode.ai/install | bash

# Or upgrade an existing install to the latest version
opencode upgrade

2. Configure a custom provider

Edit the OpenCode config file ~/.config/opencode/opencode.json (create it if it doesn't exist) and add an OpenAI-compatible provider pointing at wylon:

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "wylon": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "wylon",
      "options": {
        "baseURL": "https://api.wylon.cn/v1"
      },
      "models": {
        "moonshotai/Kimi-K2.6": { "name": "Kimi-K2.6" },
        "deepseek-ai/DeepSeek-V4-Flash": { "name": "DeepSeek-V4-Flash" }
      }
    }
  }
}

3. Provide the API Key

Set your wylon API Key as an environment variable (recommended); OpenCode reads it when calling the custom provider:

export WYLON_API_KEY=<your_wylon_API_Key>

Alternatively, after launching OpenCode you can run the /connect command, select the wylon provider, and paste your key.

4. Launch and select a model

opencode

In the UI, switch models (e.g. type /models) and pick a model under wylon, such as Kimi-K2.6, to get started.