# Using Mixlayer with Kilo Code | Fast Open-Model Inference

> Run open models like GLM and Qwen on Mixlayer's OpenAI-compatible API in Kilo Code. Setup guide for VS Code and the CLI.

- 网址：https://funcoding.ai/agents/kilo-code/ai-providers/mixlayer/
- 来源：Kilo Code 官方文档原文（英文），MIT 许可，同步于 2026-10-11
- 官方原文：https://kilo.ai/docs/ai-providers/mixlayer

---
Mixlayer is an inference platform for open models such as GLM and Qwen, with a serving stack built from scratch by core contributors to Candle. It exposes an OpenAI-compatible API and is available as a built-in provider in Kilo Code.

**Website:** [https://mixlayer.com/](https://mixlayer.com/)

## Getting an API Key

1. **Sign Up/Sign In:** Go to [Mixlayer](https://mixlayer.com/) and create an account or sign in.
2. **Navigate to API Keys:** Open the [Mixlayer console](https://console.mixlayer.com/) and go to the API Keys page.
3. **Create a Key:** Click **New Key**, give it a descriptive name (e.g., "Kilo Code"), and copy it. You will not be able to view it again.

## Configuration in Kilo Code

Mixlayer is available as a **built-in provider** in Kilo Code, so you can connect it directly — no custom provider setup needed.

**VSCode**

1. Open **Settings** (gear icon) and go to the **Providers** tab.
2. Click **Connect provider**, search for **Mixlayer**, and select it.
3. Enter your Mixlayer API key.
4. Pick a model — Kilo Code fetches the available models automatically.

**CLI**

**Method 1 — `/connect` (recommended)**

Run `kilo`, then use the `/connect` command, select **Mixlayer**, and paste your API key when prompted:

```bash
kilo
# then, inside Kilo, run:
/connect
```

**Method 2 — config file**

Set your API key and add Mixlayer in your `kilo.json` config file (`~/.config/kilo/kilo.json` or `./kilo.json`):

```bash
export MIXLAYER_API_KEY="your-api-key"
```

```jsonc
{
  "provider": {
    "mixlayer": {
      "env": ["MIXLAYER_API_KEY"],
    },
  },
  "model": "mixlayer/z-ai/glm-5.2",
}
```

## Models

Mixlayer serves open models including:

- `z-ai/glm-5.2` — 256K context
- `qwen/qwen3.5-397b-a17b` and the Qwen 3.5 / 3.6 line (vision-capable)
- `moonshotai/kimi-k2.7-code`

Tool calling and reasoning are supported across the model line. See the [Mixlayer docs](https://docs.mixlayer.com) for the full, current model list and supported parameters.

## Tips and Notes

- **Model list:** Kilo Code auto-detects available models from Mixlayer's `/v1/models` endpoint, so the picker stays current with your account.
- **Pricing:** See the [Mixlayer console](https://console.mixlayer.com/) for current per-model pricing.
- **Reasoning:** Qwen models support a thinking mode; reasoning tokens count against the output budget, so give responses enough room when reasoning is enabled.