AI Models & Providers

GitWhisper supports 9 different AI providers, giving you flexibility to choose the model that best fits your needs, budget, and preferences.

Supported Providers

OpenAI

GPT-4 models for high-quality commit messagesDefault: gpt-4o

Anthropic Claude

Advanced reasoning with Claude SonnetDefault: claude-sonnet-4-5

Google Gemini

Fast and efficient with Gemini FlashDefault: gemini-2.0-flash

xAI Grok

Grok’s latest modelsDefault: grok-2-latest

Meta Llama

Open-weight models via Llama APIDefault: llama-3-70b-instruct

DeepSeek

Cost-effective Chinese AI providerDefault: deepseek-chat

GitHub Models

AI models via GitHub’s platformDefault: gpt-4o

Ollama

Run models locally on your machineDefault: llama3.2:latest

Free (LLM7.io)

No API key required - free tierDefault: N/A

Quick Start

Using a Specific Model

Specify a model with the --model or -m flag:

# Use Claude
gw commit --model claude

# Use Gemini
gw commit -m gemini

# Use free tier (no API key needed)
gw commit --model free

Model Variants

Each provider supports multiple model variants. Specify with --model-variant or -v:

# Use GPT-4o (default OpenAI)
gw commit --model openai --model-variant gpt-4o

# Use GPT-4o mini (faster, cheaper)
gw commit --model openai --model-variant gpt-4o-mini

# Use Claude Opus (most capable)
gw commit --model claude --model-variant claude-opus-4-20250514

Provider Details

OpenAI

Overview
Setup
Variants

OpenAI’s GPT models are known for high-quality, coherent commit messages.Default Variant: gpt-4oAPI Key Required: YesAPI Endpoint: https://api.openai.com/v1/chat/completions

Get your API key from OpenAI Platform.Save it with GitWhisper:

gw save-key --model openai --key sk-proj-...

Or set as environment variable:

export OPENAI_API_KEY=sk-proj-...

gpt-4o (default) - Most capable GPT-4 model
gpt-4o-mini - Faster, more affordable
gpt-4-turbo - Previous generation
gpt-3.5-turbo - Legacy, cheapest option

Anthropic Claude

Overview
Setup
Variants

Claude excels at understanding context and writing clear, technical commit messages.Default Variant: claude-sonnet-4-5-20250929API Key Required: YesAPI Endpoint: https://api.anthropic.com/v1/messages

Get your API key from Anthropic Console.Save it with GitWhisper:

gw save-key --model claude --key sk-ant-...

Or set as environment variable:

export ANTHROPIC_API_KEY=sk-ant-...

claude-sonnet-4-5-20250929 (default) - Latest Claude Sonnet
claude-opus-4-20250514 - Most capable, slower
claude-3-5-sonnet-20241022 - Previous generation
claude-3-haiku-20240307 - Fastest, cheapest

Google Gemini

Overview
Setup
Variants

Gemini offers fast response times with good quality output.Default Variant: gemini-2.0-flashAPI Key Required: YesAPI Endpoint: https://generativelanguage.googleapis.com/v1beta/models/

Get your API key from Google AI Studio.Save it with GitWhisper:

gw save-key --model gemini --key AIza...

Or set as environment variable:

export GEMINI_API_KEY=AIza...

gemini-2.0-flash (default) - Fast and efficient
gemini-1.5-pro - More capable, slower
gemini-1.5-flash - Previous generation

xAI Grok

Overview
Setup

Grok models from xAI, known for being up-to-date and witty.Default Variant: grok-2-latestAPI Key Required: Yes

Save your API key:

gw save-key --model grok --key xai-...

Or set as environment variable:

export GROK_API_KEY=xai-...

Meta Llama

Overview
Setup

Access to Meta’s Llama models via API.Default Variant: llama-3-70b-instructAPI Key Required: Yes

gw save-key --model llama --key llama-...

Or set as environment variable:

export LLAMA_API_KEY=llama-...

DeepSeek

Overview
Setup

Cost-effective AI models with good performance.Default Variant: deepseek-chatAPI Key Required: Yes

gw save-key --model deepseek --key sk-...

Or set as environment variable:

export DEEPSEEK_API_KEY=sk-...

GitHub Models

Overview
Setup

Access AI models through GitHub’s platform.Default Variant: gpt-4oAPI Key Required: Yes (GitHub token)

Use your GitHub personal access token:

gw save-key --model github --key ghp_...

Ollama (Local)

Overview
Setup
Variants

Run AI models locally without sending data to external APIs.Default Variant: llama3.2:latestAPI Key Required: NoRequirements: Ollama must be installed and running

Install Ollama:

# macOS/Linux
curl -fsSL https://ollama.com/install.sh | sh

# Pull a model
ollama pull llama3.2

Use with GitWhisper:

gw commit --model ollama

Custom base URL:

gw set-defaults --ollama-base-url http://localhost:11434

Any model available in Ollama:

llama3.2:latest (default)
llama3.1:8b
codellama:7b
mistral:latest
phi3:latest

Specify with:

gw commit --model ollama --model-variant mistral:latest

Free Tier (LLM7.io)

Overview
Usage
When to Use

Free, anonymous access to AI models - no API key required.Powered by: LLM7.ioAPI Key Required: NoLimitations:

8,000 characters per request
60 requests/hour
10 requests/minute
1 request/second

Just use the free model:

gw commit --model free

On first use, you’ll see a disclaimer:

┌─────────────────────────────────────────────────┐
│            FREE MODEL DISCLAIMER                │
├─────────────────────────────────────────────────┤
│ This free model is powered by LLM7.io          │
│ • Your code diffs will be sent to LLM7.io      │
│ • Service availability is not guaranteed        │
│ • For production use, consider a paid provider  │
│                                                 │
│ Learn more: https://llm7.io                     │
└─────────────────────────────────────────────────┘

Do you accept these terms and wish to continue?
> yes

Setting Default Model

Set your preferred model as the default:

# Set default model
gw set-defaults --model claude

# Set default model with variant
gw set-defaults --model openai --model-variant gpt-4o-mini

# View current defaults
gw show-defaults

Once set, you can simply run:

gw commit  # Uses your default model

Model Comparison

Quality
Speed
Cost
Privacy

Based on commit message quality:

Claude (Opus/Sonnet) - Most detailed and contextual
OpenAI (GPT-4o) - Excellent balance of quality and speed
Gemini (2.0 Flash) - Good quality, very fast
Grok - Good quality with personality
DeepSeek - Solid quality, budget-friendly
Llama - Good for technical commits
GitHub Models - Similar to OpenAI
Ollama - Varies by model, privacy-focused
Free - Basic quality, rate-limited

Code Implementation

The model factory is implemented in commit_generator_factory.dart:22:

class CommitGeneratorFactory {
  static CommitGenerator create(
    String model,
    String? apiKey, {
    String? variant,
    String? baseUrl,
  }) {
    return switch (model.toLowerCase()) {
      'claude' => ClaudeGenerator(apiKey, variant: variant),
      'openai' => OpenAIGenerator(apiKey, variant: variant),
      'gemini' => GeminiGenerator(apiKey, variant: variant),
      'grok' => GrokGenerator(apiKey, variant: variant),
      'llama' => LlamaGenerator(apiKey, variant: variant),
      'deepseek' => DeepseekGenerator(apiKey, variant: variant),
      'github' => GithubGenerator(apiKey, variant: variant),
      'ollama' => OllamaGenerator(baseUrl!, apiKey, variant: variant),
      'free' => FreeGenerator(),
      _ => throw ArgumentError('Unsupported model: $model'),
    };
  }
}

Default variants are defined in model_variants.dart:10:

class ModelVariants {
  static const String openaiDefault = 'gpt-4o';
  static const String claudeDefault = 'claude-sonnet-4-5-20250929';
  static const String geminiDefault = 'gemini-2.0-flash';
  static const String grokDefault = 'grok-2-latest';
  static const String llamaDefault = 'llama-3-70b-instruct';
  static const String deepseekDefault = 'deepseek-chat';
  static const String githubDefault = 'gpt-4o';
  static const String ollamaDefault = 'llama3.2:latest';
}

Best Practices

Try Multiple Models

Different models have different strengths. Try a few to find your favorite.

Use Variants Wisely

Use faster/cheaper variants for simple commits, powerful ones for complex changes.

Go Local for Privacy

Use Ollama if you’re working with sensitive code.

Set Sensible Defaults

Configure your preferred model as default to save time.

Interactive Confirmation

Try different models during the confirmation workflow

API Key Management

Learn how to manage API keys for each provider

Configuration

Set default model and other preferences

Get Started

Core Features

Commands

Configuration

Advanced Usage

​Supported Providers

OpenAI

Anthropic Claude

Google Gemini

xAI Grok

Meta Llama

DeepSeek

GitHub Models

Ollama

Free (LLM7.io)

​Quick Start

​Using a Specific Model

​Model Variants

​Provider Details

​OpenAI

​Anthropic Claude

​Google Gemini

​xAI Grok

​Meta Llama

​DeepSeek

​GitHub Models

​Ollama (Local)

​Free Tier (LLM7.io)

​Setting Default Model

​Model Comparison

​Code Implementation

​Best Practices

Try Multiple Models

Use Variants Wisely

Go Local for Privacy

Set Sensible Defaults

​Related Features

Interactive Confirmation

API Key Management

Configuration

Build docs developers (and LLMs) love

Supported Providers

Quick Start

Using a Specific Model

Model Variants

Provider Details

OpenAI

Anthropic Claude

Google Gemini

xAI Grok

Meta Llama

DeepSeek

GitHub Models

Ollama (Local)

Free Tier (LLM7.io)

Setting Default Model

Model Comparison

Code Implementation

Best Practices

Related Features