Log in with your FastProvider API key and use all models available on your account directly inside GitHub Copilot Chat - with full tool / agent (commands) support.
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
Use your FastProvider models directly inside GitHub Copilot Chat in VS Code.
Log in once with your FastProvider API key and every model available on your
account shows up in the Copilot Chat model picker — usable in Ask, Edit and
Agent mode (with full VS Code tools / commands).
Features
🔑 Log in with your API key — stored securely in VS Code Secret Storage.
🧠 All your account models appear automatically in the model picker
(fetched live from /v1/models).
🛠️ Tool / agent support — models can call VS Code tools and run commands
in Copilot agent mode.
🔄 Refresh — re-pull the model list any time.
🚪 Log out — removes your key and hides the models.
Requirements
VS Code 1.110 or newer (Language Model Chat Provider API).
An individual GitHub Copilot plan (Free, Pro, or Pro+). The Bring Your Own
Key policy must be enabled (Business/Enterprise admins can toggle it).
A FastProvider API key (opk_live_...).
Getting started
Install the extension.
Open the Copilot Chat model picker → Manage Models → FastProvider,
or run FastProvider: Log In from the Command Palette.
Paste your API key.
Pick any FastProvider model from the model picker and start chatting.
Commands
Command
Description
FastProvider: Log In
Save your API key and load your models.
FastProvider: Log Out
Remove your key and hide the models.
FastProvider: Refresh Models
Re-query available models.
FastProvider: Manage
Quick menu (login / logout / refresh / endpoint).
Settings
Setting
Default
Description
fastprovider.baseUrl
https://www.vertexhost.eu/v1
Base URL of the FastProvider OpenAI-compatible API (must end with /v1).
Build from source
npm install
npm run compile
npx @vscode/vsce package # produces a .vsix you can install
How it works
The extension registers a LanguageModelChatProvider (vendor fastprovider).
VS Code calls provideLanguageModelChatInformation to list your models and
provideLanguageModelChatResponse to stream completions. Requests are sent to
the FastProvider OpenAI-compatible /v1/chat/completions endpoint with your key
as a Bearer token; streamed text and tool calls are translated back into VS Code
response parts so agent mode tools work end-to-end.
