GitMind: AI-Powered Commit Messages
Stop writing commit messages manually. GitMind analyzes your code changes and generates meaningful commits in seconds using 13 AI providers and 50+ models.
Why GitMind?
One Extension, 13 AI Providers
Choose from OpenAI, Anthropic, Google Gemini, DeepSeek, Grok, Perplexity, Mistral, Cohere, HuggingFace, Together AI, OpenRouter, Ollama, or GitHub Copilot. Switch anytime.
Zero Setup Options Available
- GitHub Copilot: Works instantly if you have an active subscription
- Ollama: Free, private, offline AI running locally
- Google Gemini: Industry-leading free tier (15 RPM, 2M context)
Professional Standards Built In
Generate commits following Angular, Conventional Commits, Linux Kernel, jQuery, Ember.js, Semantic Release, and more. No configuration required.
Handles Large Projects
Adaptive chunking processes massive diffs without hitting token limits. Tested with 10MB+ repositories.
Screenshots
Advanced Settings Panel (Pro Users)
|
Professional Commit Styles
|
Lifetime Pro Subscription
|
API Key Encryption (Pro Users)
|
Commit History Analysis (Pro Users)
|
Short/Verbose Message Style, Customize Prompt, Token Calculator
|
UI Settings with 13 AI Provider Support
|
Installation
Install from VS Code Marketplace
ext install ShahabBahreiniJangjoo.ai-commit-assistant
Quick Start
1. Pick Your AI Provider
| Provider |
Best For |
Free Tier |
Setup Time |
| GitHub Copilot |
Zero setup, existing subscribers |
No |
0 min |
| Google Gemini |
Best free tier (15 RPM, 2M context) |
Yes |
2 min |
| Ollama |
Complete privacy, offline usage |
Unlimited |
5 min |
| DeepSeek |
Advanced reasoning, cost-effective |
50 RPM |
2 min |
| OpenAI |
Industry standard, multimodal |
No |
2 min |
| Anthropic |
Superior reasoning, long context |
No |
2 min |
| Grok |
Real-time information, X.AI models |
No |
2 min |
| Perplexity |
Web search capabilities, real-time data |
No |
2 min |
| Mistral |
European AI, multilingual support |
No |
2 min |
| Cohere |
Enterprise-grade, multilingual |
Trial |
2 min |
| Together AI |
Open-source models, cost-effective |
Trial |
2 min |
| OpenRouter |
Multi-model access, unified API |
Trial |
2 min |
| Hugging Face |
Custom models, research access |
Trial |
2 min |
- Open VS Code Source Control panel
- Click settings icon in GitMind section
- Select provider
- Add API key (not needed for GitHub Copilot or Ollama)
- Choose model
3. Generate
- Stage changes
- Click AI button in Source Control
- Review generated message
- Commit
AI Provider Support
All Supported Providers
| Provider |
Featured Models |
Context |
API Setup |
| GitHub Copilot |
gpt-4o, claude-3.5-sonnet, o3 |
128k |
VS Code Copilot |
| Google Gemini |
2.5-flash, 2.5-pro, 2.0-flash |
2M |
AI Studio |
| Grok (X.ai) |
grok-3, grok-3-fast, grok-3-mini |
128k |
X.ai Console |
| DeepSeek |
reasoner, chat |
128k |
DeepSeek Platform |
| Perplexity |
sonar-pro, sonar-reasoning, sonar |
127k |
Perplexity Settings |
| Mistral AI |
large-latest, medium, small |
128k |
Mistral Console |
| Ollama |
deepseek-r1, llama3.3, phi4, qwen3 |
128k |
Ollama Download |
| OpenAI |
gpt-4o, gpt-4.1, o3, o4-mini |
128k |
OpenAI Platform |
| Anthropic |
claude-opus-4, sonnet-4, haiku |
200k |
Anthropic Console |
| Together AI |
Llama-3.3-70B, Mixtral-8x7B |
128k |
Together Platform |
| Hugging Face |
Mistral-7B, Zephyr-7B, OpenHermes |
32k |
HF Token |
| Cohere |
command-r, command-a-03-2025 |
128k |
Cohere Dashboard |
| OpenRouter |
Multiple providers & models |
Varies |
OpenRouter Keys |
Free vs Pro
| Feature |
Free |
Pro |
| AI Providers |
13 providers (all) |
13 providers (all) |
| Models |
50+ models |
50+ models |
| Commit Styles |
Basic, Conventional |
11 professional styles |
| Git Integration |
Yes |
Yes |
| Multi-Repository Support |
Yes |
Yes |
| Diagnostics & Token Estimation |
Yes |
Yes |
| Verbose/Concise Messages |
Yes |
Yes |
| Prompt Customization |
Yes |
Yes + Save Last Prompt |
| API Key Storage |
Plain text |
Encrypted (SecretStorage) |
| Large Diff Processing |
Limited |
Token-aware chunking |
| Commit Body Lines |
Fixed (5 lines) |
Configurable (3-15) |
| Summary Length |
Fixed |
Configurable (50-100 chars) |
| Commit History Analysis |
✗ |
✓ |
| Changelog Generation |
✗ |
✓ |
| Gitmoji Support |
✗ |
✓ |
| Custom API Endpoints |
✗ |
✓ |
| Multi-Device License |
Single device |
Multiple devices |
Features
Core (Free)
Multi-Provider AI
Access 13 providers with unified configuration: OpenAI, Anthropic, Google Gemini, DeepSeek, Grok, Perplexity, Mistral, Cohere, HuggingFace, Together AI, OpenRouter, Ollama, GitHub Copilot. Switch providers instantly.
Basic Commit Style
Simple, straightforward commit messages without prefixes or conventions.
Smart Diff Analysis
Automatic staging detection, binary file handling, merge conflict awareness.
Repository Integration
Native VS Code SCM panel. Multi-repository workspace support with independent buttons per repository.
Diagnostics
Token estimation, provider/model summary, request preview.
Pro Features
Professional Commit Styles
10 industry-standard formats:
- Conventional Commits: Semantic versioning support with types and scopes
- Angular: Enterprise-grade with strict conventions
- Ember.js: Tag-based semantic categorization
- EmojiGit: Visual semantic commits with custom emojis
- Gitmoji: Official gitmoji.dev specification
- Semantic Release: Automated release-optimized commits
- Commitizen: Interactive guided commits with validation
- Karma (Google): Google's strict enterprise convention
- Linux Kernel: Traditional kernel development convention
- jQuery: JavaScript project convention with issue tracking
Changelog Generation
AI-powered changelog from git history with:
- 3-tier version detection (git tags, commit messages, package.json)
- Existing changelog policy awareness (format, categories, emoji usage preserved)
- Keep a Changelog specification compliance
- Configurable commit analysis (10-500 commits, default 100)
- Group by version tags option
- Create new, update existing, or preview modes
Commit History Analysis
Learn from repository patterns:
- Analyzes 10-500 past commits (default 50)
- Optional author and date information
- AI-powered pattern recognition
- Markdown report generation
- Configurable analysis depth
Large Diff Processing
Token-aware adaptive chunking with:
- Hunk-aware intelligent splitting
- Automatic split/merge for massive diffs
- Configurable concurrency (1-8 workers, default 3)
- Retry logic with exponential backoff (0-5 retries, default 2)
- Retryable error detection (rate limits, timeouts, network errors)
- Progress reporting with real-time updates
API Key Encryption
Secure storage using VS Code SecretStorage API:
- Keys encrypted at rest
- Inaccessible to other extensions
- Automatic migration from plaintext
- Pro-only feature
Advanced Customization
- Custom commit body line limits (3-15 lines)
- Custom summary length limits (50-100 characters)
- Save/reuse custom prompts
- Gitmoji support with placement control (summary, body, or both)
- Custom emoji mappings for commit types
- Custom API provider endpoints
Multi-Device License
Use across multiple development machines with subscription-based access.
Configuration
Access via Command Palette: GitMind: Open Settings
Provider Settings
- AI provider selection
- Secure API key configuration
- Model selection
Message Formatting
- Commit style (11 professional formats in Pro)
- Verbosity control (verbose/concise)
- Custom scope and type
Pro Settings
- Encrypted API key storage
- Commit history analysis
- Custom body/summary limits
- Gitmoji configuration
- Changelog generation
Privacy
GitMind collects anonymous usage data (provider statistics, error reports, usage analytics). No code content, file names, personal information, API keys, or repository details are collected.
Disable telemetry:
- Open VS Code Settings (Ctrl/Cmd + ,)
- Search "telemetry"
- Set "Telemetry: Telemetry Level" to "off"
Requirements
- Visual Studio Code ^1.100.0
- API key from chosen provider OR Ollama for local deployment
Support
License
MIT License - see LICENSE for details