Guide
Guide

Configuration

Wingman has a built-in configuration window that will persist configuration locally into a settings file on your machine.

Settings can be accessed from the side panel, or under the Wingman option in the terminal pane.

Storage

The Wingman storage location is shown below:

Storage Location:

/Home Directory/.wingman

Example on macOS:

/Users/username/.wingman

Supported Models

We aim to support the best models available. We allow the user to configure separate models for chat and code completion, this is especially helpful when running AI models locally. Here is a list of which models we support for each provider:

Anthropic

You can use the following models:

  • Claude 3.7 Sonnet
  • Claude 3.5 Haiku
  • Claude 3.5 Sonnet
  • Claude 3 Opus
NOTE

Sonnet 3.7 is a pretty aggressive model, you can enable thinking mode (2048 token budget) or spark mode which will remove specific focus constraints on 3.7. Spark mode is the raw model output without the prompt restricting it.

Anthropic prompt caching is used for optimization reasons

NOTE - Unlike using Ollama, your data is not private and will not be sanitized prior to being sent to Anthropic.

OpenAI

You can use the following models:

  • o1
  • o3-mini (medium reasoning)
  • GPT-4o
  • GPT-4o-mini
  • GPT-4-Turbo

NOTE - Unlike using Ollama, your data is not private and will not be sanitized prior to being sent to OpenAI

Google AI Studio

You can use the following models:

  • Gemini 2.5 Pro
  • Gemini 2.5 Flash
NOTE

Gemini 2.5 Pro supports image generation capabilities, allowing you to create visual content directly within your workflow. This can be particularly useful for generating UI mockups, diagrams, or other visual assets.

NOTE - Unlike using Ollama, your data is not private and will not be sanitized prior to being sent to Google AI.

AzureAI

NOTE - AzureAI has general latency due to content filters on models by default. This can cause delays in responses and may require additional configuration to disable content filters.

You can use the following models:

  • o1
  • o3-mini (medium reasoning)
  • GPT-4o
  • GPT-4o-mini
  • GPT-4-Turbo
  • GPT4
  • GPT-1o

NOTE - Unlike using Ollama, your data is not private and will not be sanitized prior to being sent Azure OpenAI.

OpenRouter

OpenRouter provides access to a wide range of models from different providers through a unified API.

You can use models from:

  • Anthropic
  • OpenAI
  • Google
  • Mistral
  • And many other providers

NOTE - OpenRouter acts as a proxy to these models, so your data is not private and will not be sanitized prior to being sent to the respective model providers.

Ollama

Wingman uses a full agentic flow. You must use an Ollama model that supports tools. If you would like to add a model, please open an issue in Github.

NOTE - You can use any quantization for a supported model, you are not limited.

Example: deepseek-coder:6.7b-instruct-q4_0

Supported Models for Code Completion:

Supported Models for Chat:

LMStudio

LMStudio provides a user-friendly interface for running local AI models with an OpenAI-compatible API.

Supported Models:

  • Any model that can be loaded into LMStudio and exposed via its API server
  • Particularly good with Mistral, Llama, and CodeLlama variants

NOTE - When using LMStudio, your data remains on your local machine, providing enhanced privacy compared to cloud-based solutions.

Hugging Face

NOTE - These are out of date

Supported Models for Code Completion:

Supported Models for Chat:

NOTE - Unlike using Ollama, your data is not private and will not be sanitized prior to being sent.

Settings

Settings for the extension are broken down into 4 categories.

  • AI Provider
  • Interaction Settings
  • Agent Features

AI Provider

Provider settings will include which model to use, endpoints and the API key to use. These will save per provider allowing you to switch on the fly.

Interaction Settings

General extension settings are persisted separately from the AI provider, here is a breakdown of the general settings:

Code completion enabled

Code completion can run automatically triggered by line returns, spaces and tabs. Or can by hotkeying the "Wingman: Code Complete" command.

Code streaming

This is an experimental version of code complete that attempts to return results faster, allowing the user to see incremental changes as they accept.

Code context window

During code completion, this controls the amount of surrounding text passed to the AI provider, giving better auto completion results.

Code max tokens

The maximum amount of tokens the code models can generate during code completion.

Chat max tokens

Controls the maximum about of tokens the AI provider will return.

Agent Features

Agent Features are general features for the overall experience.

Vibe Mode

Automatically accept commands and proposed file changes. When off, you will be prompted to accept each command and file change - you can view the diff before accepting. If you reject, you will be asked what modifications you'd like to make.

Auto Fix

Automatically fix Linting/Import errors for any file modified by Wingman.

Audio Alert

Emits an audio cue when the session has completed, or when in manual mode, emits an audio cue for each command or file you need to accept or reject.