Wingman AI

This extension integrates the Wingman AI platform into VS Code's Copilot Chat by configuring the built-in "Custom Endpoint" language model provider that ships with VS Code (1.123+). It auto-discovers the models available on your Wingman backend and registers them as a "Wingman" language model group — VS Code itself handles all chat traffic (OpenAI Responses API, streaming, tool calling, thinking, images).

⚠️ Important Notice: This extension requires the Wingman AI Platform and should not be installed unless you are running a compatible Wingman backend.

How it works

On first startup (and whenever you run Wingman: Sync Models), the extension:

1. Queries GET {baseUrl}/models on your Wingman backend
2. Matches the reported model IDs against its model catalog (names, token limits, capabilities, reasoning effort levels)
3. Registers a Wingman provider group with VS Code's built-in customendpoint vendor (apiType: responses, zeroDataRetentionEnabled: true — requests are stateless, no previous_response_id is ever sent)

After that, the models appear in the Copilot Chat model picker. The resulting configuration lives in your VS Code profile's chatLanguageModels.json (command: Open Language Models File) and can be edited or removed there — re-running Wingman: Sync Models after removal re-creates it.

Features

• Native Copilot Chat integration — uses VS Code's built-in OpenAI-compatible provider; no custom wire protocol code
• Multi-vendor model support — automatically discovers available models from your backend
• Zero Data Retention — Responses API in stateless mode (store: false, no previous_response_id)
• Tool calling, vision, thinking — capabilities forwarded per model
• Reasoning controls — per-model "Thinking Effort" picker where supported

Supported Models

The extension auto-discovers models from your backend. The following model families are supported:

Chat Model	Model IDs
GPT 5.5	gpt-5.5
GPT 5.4	gpt-5.4
GPT 5.4 mini	gpt-5.4-mini
GPT 5.2	gpt-5.2
GPT 5.1	gpt-5.1
Codex 5.3	gpt-5.3-codex
Codex 5.2	gpt-5.2-codex
Gemini 3.5 Flash	gemini-3.5-flash
Gemini 3.1 Pro	gemini-3.1-pro, gemini-3.1-pro-preview
Gemini 3 Pro	gemini-3-pro, gemini-3-pro-preview
Gemini 3 Flash	gemini-3-flash, gemini-3-flash-preview
Opus 4.8	claude-opus-4-8
Opus 4.7	claude-opus-4-7
Opus 4.6	claude-opus-4-6
Opus 4.5	claude-opus-4-5
Sonnet 4.6	claude-sonnet-4-6
Sonnet 4.5	claude-sonnet-4-5
Haiku 4.6	claude-haiku-4-6
Haiku 4.5	claude-haiku-4-5
Devstral Medium	devstral-medium, devstral-medium-latest, devstral-latest, devstral
Devstral Small	devstral-small, devstral-small-latest
GLM 5.1	glm-5.1
GLM 5	glm-5
GLM 4.7	glm-4.7
GLM 4.7 Flash	glm-4.7-flash
Qwen 3.7 Max	qwen3.7-max
Qwen 3.6	qwen3.6-plus, qwen3.6
Qwen 3.6 Flash	qwen3.6-flash
Qwen 3.5	qwen3.5-plus, qwen3.5
Qwen 3	qwen3-next, qwen3
Qwen 3 Coder	qwen3-coder-plus, qwen3-coder-flash, qwen3-coder-next, qwen3-coder

The extension picks the first available model from each group based on what your backend reports.

Configuration

Setting	Description	Default
wingman.baseUrl	Base URL of your Wingman API	http://localhost:4242/v1
wingman.apiKey	API key for authentication	-

Settings are read when syncing. To apply changes, remove the existing "Wingman" group (Open Language Models File) and run Wingman: Sync Models again.

Requirements

• VS Code 1.123.0 or later (built-in Copilot Chat with the "Custom Endpoint" provider)
• A running Wingman AI

License

MIT