Start with 1: run Copilot on AI Reserve models, so every Copilot chat is metered on your AI Reserve key. 2 is optional on top: the AI Reserve assistant for account questions inside Copilot Chat.
1. Run Copilot on AI Reserve models
Start here. Our models in Copilot's own model picker, for Chat, agent mode and edits, made the default so nobody stays on GitHub's model. Your AI Reserve key is what gets debited.
2. Optional: the AI Reserve assistant in Copilot Chat
Account tools inside Copilot Chat: budget, usage, prices, keys, or a second opinion from a named model. Sign in, nothing to paste.
Any AI Reserve key works
Paste a key you already have from API Keys, or create one there. Whatever that key belongs to, you, your team or your organization, is what pays, with its spend caps and model rules applied. GitHub bills nothing extra, and these requests do not count against Copilot premium-request allowances.
For yourself: install the AI Reserve extension
Our models sit under the AI Reserve heading in the picker and every detail line starts with “AI Reserve”, so they are never confused with GitHub's copies. The Marketplace listing is on its way; this page switches to a one-click install the day it is live.
For your whole organization: your GitHub admin adds one key
On GitHub Enterprise with Copilot Business or Enterprise seats, an admin adds AI Reserve once as a custom model provider. Our models then appear in every developer's picker, including Copilot Chat on github.com, with nothing to install.
https://api.aireserve.com/v1
claude-sonnet-4-6, claude-opus-4-8, gpt-5.5, gpt-5.4-mini, gemini-2.5-pro
A curated coding set. Any chat model id from the catalog works; add auto to let AI Reserve pick per request.
API key: the organization key from your AI Reserve administrator (created with the button above when you are one). Enable the models for your organizations and save; they appear within a few minutes. When AI Reserve adds a model, add its id to this list; auto and the extension pick new models up on their own.
ChatDefaultModel policy with the value aireserve (VS Code 1.127+, by MDM or group policy like any VS Code policy). Every managed machine then starts new chats on AI Reserve Auto and users cannot override it.Together they guarantee every chat, agent and edit request is metered on your AI Reserve key. Inline ghost-text completions keep working as before: they stay on GitHub's models either way. No admin yet? Each developer can pin AI Reserve models to the top of the picker (hover a model → pin) and hide GitHub's under Manage Models (hover → eye icon).
Have the aireserve CLI? aireserve copilot sync writes both of these for you and keeps them current.
aireserve copilot sync
Reuses the key aireserve configure stored; add --dry-run to preview the changes first.
By hand instead? VS Code: run Chat: Manage Language Models → Add Models → Custom Endpoint, name it AI Reserve, and replace the generated entry with this. VS Code prompts for the key once and keeps it in secret storage, so the file never holds a secret:
[
{
"name": "AI Reserve",
"vendor": "customendpoint",
"apiKey": "${input:aireserveApiKey}",
"apiType": "chat-completions",
"models": [
{
"id": "auto",
"name": "AI Reserve · Auto",
"url": "https://api.aireserve.com/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 128000,
"maxOutputTokens": 16000
},
{
"id": "claude-sonnet-4-6",
"name": "AI Reserve · claude-sonnet-4-6",
"url": "https://api.aireserve.com/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 1000000,
"maxOutputTokens": 128000
},
{
"id": "claude-opus-4-8",
"name": "AI Reserve · claude-opus-4-8",
"url": "https://api.aireserve.com/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 1000000,
"maxOutputTokens": 128000
},
{
"id": "gpt-5.5",
"name": "AI Reserve · gpt-5.5",
"url": "https://api.aireserve.com/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 1050000,
"maxOutputTokens": 128000
},
{
"id": "gpt-5.4-mini",
"name": "AI Reserve · gpt-5.4-mini",
"url": "https://api.aireserve.com/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 400000,
"maxOutputTokens": 128000
},
{
"id": "gemini-2.5-pro",
"name": "AI Reserve · gemini-2.5-pro",
"url": "https://api.aireserve.com/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 1048576,
"maxOutputTokens": 65536
}
]
}
]Add any other catalog id as another entry with the same url.
Copilot CLI: set these in your shell profile.
export COPILOT_PROVIDER_BASE_URL="https://api.aireserve.com/v1" export COPILOT_PROVIDER_TYPE="openai" export COPILOT_PROVIDER_API_KEY="<your aireserve_api_… key>" export COPILOT_MODEL="auto"
Optional, on top of section 1. Tools inside the Copilot you already use: check your team's usage and budget, see what models cost, manage API keys, or ask a named model for a second opinion (“using fable 5.1 ping me”). You sign in with your portal account; there is no key to paste, and each person only sees what they can already see in the portal. With an AI Reserve model selected in the picker, the assistant runs entirely on AI Reserve too.
Copilot Business or Enterprise: your GitHub admin turns on the MCP servers in Copilot policy once (off by default). Copilot Pro and Free need nothing.
VS Code
Click Install, then Allow when VS Code asks to sign in to AI Reserve.
Open Copilot Chat in Agent mode, click the tools icon (Configure your MCP server) → Add MCP Tools, paste this, and click Allow on the sign-in prompt:
{
"servers": {
"aireserve": {
"type": "http",
"url": "https://api.aireserve.com/mcp"
}
}
}copilot mcp add --transport http aireserve https://api.aireserve.com/mcp
First use opens the AI Reserve sign-in in your browser (Copilot CLI 1.0.11 or newer). The GitHub Copilot desktop app picks up the CLI's servers automatically.
In the app itself: Customize → MCP → Add custom server, type HTTP, URL https://api.aireserve.com/mcp, no headers, sign in when prompted.
These run without a browser, so a repository admin gives them a token instead of a sign-in:
COPILOT_MCP_AIRESERVE_TOKEN (repository Settings → Environments → copilot).{
"mcpServers": {
"aireserve": {
"type": "http",
"url": "https://api.aireserve.com/mcp",
"headers": {
"Authorization": "Bearer $COPILOT_MCP_AIRESERVE_TOKEN"
},
"tools": [
"whoami",
"list-models",
"get-model",
"get-connection-guide",
"send-message"
]
}
}
}Tools are listed explicitly because the agent calls them without asking you first.
Try it
In Copilot Chat, switch to Agent mode and type:
using fable 5.1 ping me
The reply ends with the exact model that answered (Model: claude-fable-5-1). The same loop is one command, /mcp.aireserve.ask-model. Also try “which models can I use and what do they cost”, “what did my team spend this week”, “how much budget do I have left”, “list my API keys”.
Copilot Chat on github.com gets our models through the organization-wide setup in section 1; that is the only way any vendor reaches its picker, and the assistant cannot be added there (GitHub wires only its own MCP server into that chat).
The Copilot cloud agent and code review also run on github.com and take the assistant through the repository configuration in section 2.
Every request is metered on the AI Reserve key or account it came in on and billed to that organization's wallet. Model access and spend caps are enforced server-side: an administrator can block models or cap spend regardless of what a developer selects, and revoking a key or a person's access applies on the next request.
Keys pasted into VS Code live in its secret storage, never in settings or files. The assistant signs each developer in individually, so it can only see and do what that person already can in the portal. Nothing runs on the developer's machine; the endpoints are https://api.aireserve.com/v1 and https://api.aireserve.com/mcp.
Need a key? Any key from your API Keys page works; organization keys are created by your AI Reserve administrator. Trouble connecting? Ask your administrator to confirm your account is active, or see the developer documentation for the full GitHub Copilot guide.