Skip to main content
A VS Code extension that integrates Vercel AI Gateway models directly into VS Code’s native chat interface, enabling seamless AI-powered development workflows.

​
Features

Multiple AI Models

Access all Vercel AI Gateway models including GPT-4o, Claude, Gemini, and more

Native Integration

Works seamlessly with VS Code’s built-in chat interface and Copilot

Streaming Responses

Real-time streaming of AI responses for instant feedback

Tool Calling

Full support for VS Code tool integration and function calling

​
Installation

Install the extension from the VS Code Marketplace:
  1. Open VS Code
  2. Go to Extensions (Cmd/Ctrl + Shift + X)
  3. Search for “Vercel AI Gateway”
  4. Click Install
Or install via the command line: 
code --install-extension SferaDev.vscode-extension-vercel-ai

​
Authentication

The extension supports two authentication methods:

​
API Key Authentication

The simplest way to authenticate is with a Vercel AI Gateway API key:
  1. Open Command Palette (Cmd/Ctrl + Shift + P)
  2. Search for “Vercel AI Gateway: Manage Authentication”
  3. Select “API Key”
  4. Enter your API key (starts with vck_)
You can obtain an API key from the Vercel AI Gateway dashboard.

​
OIDC Authentication (Vercel CLI)

For enhanced security, you can use OIDC authentication via the Vercel CLI:
  1. Ensure you’re logged in to the Vercel CLI (vercel login)
  2. Open Command Palette (Cmd/Ctrl + Shift + P)
  3. Search for “Vercel AI Gateway: Manage Authentication”
  4. Select “Vercel OIDC”
  5. Choose your team and project
OIDC tokens are automatically refreshed when they expire.

​
Usage

Once authenticated, the extension automatically registers as a language model provider in VS Code.

​
Using with VS Code Chat

  1. Open the Chat panel (Cmd/Ctrl + Shift + I)
  2. Click the model selector dropdown
  3. Choose a model from “Vercel AI Gateway”
  4. Start chatting!

​
Using with GitHub Copilot Chat

The extension integrates with GitHub Copilot Chat, allowing you to use Vercel AI Gateway models as an alternative to the default Copilot models.

​
Available Models

The extension dynamically fetches available models from the Vercel AI Gateway. Supported models include:
ProviderModels
OpenAIGPT-4o, GPT-4o Mini, GPT-4 Turbo
AnthropicClaude 3.5 Sonnet, Claude 3 Opus, Claude 3 Haiku
GoogleGemini 1.5 Pro, Gemini 1.5 Flash
And more…All models supported by Vercel AI Gateway
All models support:
  • Text generation with streaming
  • Tool calling (where supported by the underlying model)
  • Token counting

​
Session Management

The extension supports multiple authentication sessions, allowing you to switch between different API keys or OIDC configurations.

​
Managing Sessions

  1. Open Command Palette (Cmd/Ctrl + Shift + P)
  2. Search for “Vercel AI Gateway: Manage Authentication”
  3. Choose from:
    • Add new authentication - Add a new API key or OIDC session
    • Switch active session - Switch between existing sessions
    • Remove session - Delete a saved session

​
Troubleshooting

  1. Ensure you’re authenticated (check via “Manage Authentication” command)
  2. Try reloading VS Code (Cmd/Ctrl + Shift + P → “Reload Window”)
  3. Check the Output panel for error messages (select “Vercel AI Gateway” from the dropdown)
  1. Verify you’re logged in to the Vercel CLI: vercel whoami
  2. If not logged in, run: vercel login
  3. Ensure the project you selected has AI Gateway enabled
  1. Verify your API key starts with vck_
  2. Check that your API key has the necessary permissions
  3. Try removing and re-adding the session

​
Development

To contribute or run the extension locally: 
# Clone the repository
git clone https://github.com/SferaDev/SferaDev.git
cd SferaDev

# Install dependencies
pnpm install

# Build the extension
pnpm --filter vscode-extension-vercel-ai build

# Open in VS Code for debugging
code apps/vscode-ai-gateway
Press F5 to launch the Extension Development Host.

​
Source Code

The extension is part of the SferaDev monorepo and can be found at: 
apps/vscode-ai-gateway/