---
title: Adding AI to VS Code using Continue and Generative APIs
description: Learn how to integrate AI-powered code models into VS Code with Continue and Scaleway's Generative APIs.
tags: generative-apis ai machine-learning language-models code-assistance vs-code continue
dates:
validation: 2025-08-20
posted: 2025-02-14
---
import Requirements from '@macros/iam/requirements.mdx'
AI-powered coding is transforming software development by automating repetitive tasks, generating code, improving code quality, and even detecting and fixing bugs. By integrating AI-driven tools, developers can significantly boost productivity and streamline their workflows.
This guide provides a step-by-step guide on how to integrate AI-powered code models into VS Code using Continue and Scaleway's Generative APIs.
- A Scaleway account logged into the [console](https://console.scaleway.com)
- [Owner](/iam/concepts/#owner) status or [IAM permissions](/iam/concepts/#permission) allowing you to perform actions in the intended Organization
- A valid [API key](/iam/how-to/create-api-keys/) for API authentication
- Installed [Visual Studio Code](https://code.visualstudio.com/) on your local machine
## Install Continue in VS Code
You can install Continue directly from the [Visual Studio Marketplace](https://marketplace.visualstudio.com/items?itemName=Continue.continue) or via the command line:
```bash
code --install-extension continue.continue
```
## Configure Continue to use Scaleway’s Generative APIs
### Configure Continue through the graphical interface
To link Continue with Scaleway's Generative APIs, you can use built-in menus from Continue in VS Code.
- Click **Continue** in the menu on the left.
- In the prompt section, click on **Select model** dropdown, then on **Add Chat model**.
- Select **Scaleway** as provider.
- Select the model you want to use (we recommend `Qwen 3.5 397B` to get started with agentic coding tasks).
- Enter your **Scaleway secret key**.
To start with, we recommend you use a Scaleway secret key having access to your `default` Scaleway project.
These actions will automatically edit your `config.yaml` file. To edit it manually, see [Configure Continue through configuration file](#configure-continue-through-a-configuration-file).
Agents, embeddings, and autocomplete models are not yet supported through graphical interface configuration. Manually edit the configuration to enable them. See [Configure Continue through configuration file](#configure-continue-through-configuration-file) for more information.
### Configure Continue through a configuration file
To link Continue with Scaleway's Generative APIs, you can configure a settings file:
- Open your `config.yaml` settings file:
- If you have already configured a **Local Assistant**, click **Local Assistant**, then click the **wheel icon** to open your existing `config.yaml`
- Otherwise, create a `config.yaml` file inside your `.continue` directory.
- Add the following configuration to enable Scaleway's Generative API. This configuration uses three different models for each tasks:
- `qwen3.5-397b-a17b` for agentic workflows through a chat interface
- `qwen3.6-35b-a3b` for autocompletion when editing a file
- `qwen3-embedding-8b` for embedding and retrieving code context
```yaml
name: Continue Config
version: 0.0.1
models:
- name: Qwen3.5 397b - Scaleway
provider: openai
model: qwen3.5-397b-a17b
apiBase: https://api.scaleway.ai/v1/
apiKey: ###SCW_SECRET_KEY###
defaultCompletionOptions:
maxTokens: 4000
contextLength: 40000
roles:
- chat
- apply
- embed
- edit
capabilities:
- tool_use
- name: Autocomplete - Scaleway
provider: openai
model: qwen3.6-35b-a3b
apiBase: https://api.scaleway.ai/v1/
apiKey: ###SCW_SECRET_KEY###
defaultCompletionOptions:
maxTokens: 8000
contextLength: 100000
roles:
- autocomplete
- name: Embeddings Model - Scaleway
provider: openai
model: qwen3-embedding-8b
apiBase: https://api.scaleway.ai/v1/
apiKey: ###SCW_SECRET_KEY###
roles:
- embed
embedOptions:
maxChunkSize: 256
maxBatchSize: 32
context:
- provider: problems
- provider: tree
- provider: url
- provider: search
- provider: folder
- provider: codebase
- provider: web
params:
n: 3
- provider: open
params:
onlyPinned: true
- provider: docs
- provider: terminal
- provider: code
- provider: diff
- provider: currentFile
```
- Save the file at the correct location:
- Linux/macOS: `~/.continue/config.yaml`
- Windows: `%USERPROFILE%\.continue\config.yaml`
- In **Local Assistant**, click on **Reload config** or restart VS Code.
Alternatively, a `config.json` file can be used with the following format. Note that this format is deprecated, and we recommend using `config.yaml` instead.
```json
{
"models": [
{
"model": "qwen3.5-397b-a17b",
"title": "Qwen3.5 397B - Scaleway",
"provider": "openai",
"apiKey": "###SCW_SECRET_KEY###"
}
],
"embeddingsProvider": {
"model": "qwen3-embedding-8b",
"provider": "openai",
"apiKey": "###SCW_SECRET_KEY###"
},
"tabAutocompleteModel": {
"model": "qwen3.6-35b-a3b",
"title": "Autocomplete - Scaleway",
"provider": "openai",
"apiKey": "###SCW_SECRET_KEY###"
}
}
```
For more details on configuring `config.yaml`, refer to the [official Continue documentation](https://docs.continue.dev/reference).
If you want to limit access to a specific Scaleway Project, you should add the field `"apiBase": "https://api.scaleway.ai/###PROJECT_ID###/v1/"` for each model (ie. `models`, `embeddingsProvider` and `tabAutocompleteModel`) since the default URL `https://api.scaleway.ai/v1/` can only be used with the `default` project.
## Activate Continue in VS Code
After configuring the API, open VS Code and activate Continue:
- Open the **Command Palette** (`Ctrl+Shift+P` on Windows/Linux, `Cmd+Shift+P` on Mac)
- Type `"Continue"` and select the appropriate command to enable AI-powered assistance.
Enabling tab completion **may lead to higher token consumption** as the model generates predictions for every keystroke. Be mindful of your API usage and adjust settings accordingly to avoid unexpected costs. For more information, refer to the [official Continue documentation](https://docs.continue.dev/reference#tabautocompleteoptions).
## Going further
You can add more parameters to configure your model's behavior by editing `config.yaml`.
For instance, you can add the following `chatOptions.baseSystemMessage` value to modify LLM messages `"role":"system"` and/or `"role":"developer"` and provide less verbose answers:
```yaml
model:...
chatOptions:
baseSystemMessage: "You are an expert developer. Only write concise answers."
```