Kilo Code and OpenCode

Kilo Code and OpenCode can connect to Inceptron through their OpenAI-compatible provider settings.

Install Kilo Code

Install the Kilo Code extension in your editor, or install the Kilo Code CLI if you prefer using it from a terminal.

Add Inceptron as a Provider

In Kilo Code, add a new OpenAI-compatible provider with these settings:

Setting	Value
Provider name	Inceptron
Base URL	https://api.inceptron.io/v1
API key	Your Inceptron API key

You can create and manage API keys from your Inceptron account dashboard.

Configure the Provider

note

Most supported models are listed on Models.dev, which Kilo Code and OpenCode use for model discovery and metadata. If the Inceptron model you want is already available through the model picker, a manual model definition may not be necessary. New models can take some time to appear on Models.dev; the explicit custom-provider configuration below lets you use them in the meantime or override their metadata.

Set your API key in the environment so that it is not stored directly in the configuration file:

export INCEPTRON_API_KEY="your-api-key"

Kilo Code

Add the following to kilo.jsonc in your project root, or to ~/.config/kilo/kilo.jsonc to make Inceptron available in every project:

{
  "$schema": "https://app.kilo.ai/config.json",
  "model": "inceptron/zai-org/GLM-5.3",
  "provider": {
    "inceptron": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Inceptron",
      "options": {
        "baseURL": "https://api.inceptron.io/v1",
        "apiKey": "{env:INCEPTRON_API_KEY}"
      },
      "models": {
        "zai-org/GLM-5.3": {
          "name": "GLM-5.3",
          "reasoning": true,
          "tool_call": true,
          "limit": {
            "context": 1048576,
            "output": 1048576
          },
          "interleaved": {
            "field": "reasoning_content"
          },
          "options": {
            "reasoningEffort": "high"
          },
          "variants": {
            "low": { "reasoningEffort": "low" },
            "high": { "reasoningEffort": "high" },
            "max": { "reasoningEffort": "max" }
          }
        }
      }
    }
  }
}

OpenCode

OpenCode uses the same provider and model settings. Add this configuration to opencode.jsonc in your project root, or to ~/.config/opencode/opencode.jsonc for a global configuration:

{
  "$schema": "https://opencode.ai/config.json",
  "model": "inceptron/zai-org/GLM-5.3",
  "provider": {
    "inceptron": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Inceptron",
      "options": {
        "baseURL": "https://api.inceptron.io/v1",
        "apiKey": "{env:INCEPTRON_API_KEY}"
      },
      "models": {
        "zai-org/GLM-5.3": {
          "name": "GLM-5.3",
          "reasoning": true,
          "tool_call": true,
          "limit": {
            "context": 1048576,
            "output": 1048576
          },
          "interleaved": {
            "field": "reasoning_content"
          },
          "options": {
            "reasoningEffort": "high"
          },
          "variants": {
            "low": { "reasoningEffort": "low" },
            "high": { "reasoningEffort": "high" },
            "max": { "reasoningEffort": "max" }
          }
        }
      }
    }
  }
}

GLM-5.3 supports a 1M-token context window (1,048,576 tokens). The model-level reasoningEffort sets the default. GLM-5.3 always uses reasoning; select or cycle the low, high, and max variants to change the effort for the current task. Higher efforts allow the model to spend more work on complex tasks. The interleaved.field setting preserves reasoning_content between turns.

Select a Model

Select one of the currently supported Inceptron model IDs:

Model ID
deepseek-ai/DeepSeek-V4-Flash-0731
moonshotai/Kimi-K2.7-Code
moonshotai/Kimi-K2.6
zai-org/GLM-5.3
zai-org/GLM-5.2

After saving the provider, start a Kilo Code or OpenCode session and choose the Inceptron provider with the model you want to use.