# AII β AI-Powered Terminal Assistant
**AII** (pronounced *"eye"*) is your **intelligent command-line companion** that brings Claude, GPT, and Gemini directly into your terminal through natural language commands.
**π Version 0.4.7** β Template Library & Analytics | **v0.4.5** β Git PR & Branch | **v0.4.4** β Streaming + Output modes
## Why AII?
**One CLI. 25 AI Functions. Natural Language. Zero Configuration Headaches.**
Stop context-switching between your terminal and ChatGPT. Get instant AI assistance for git commits, code generation, translation, content writing, shell automation, and analysisβall without leaving your terminal.
**Perfect for:**
- **Developers**: AI-powered git commits, code generation, and shell automation
- **Creators**: Content writing, translation, and social media posts
- **Professionals**: Email templates, document analysis, and research assistance
Built with the Pydantic AI framework for reliable, multi-provider LLM integration (Anthropic Claude, OpenAI GPT, Google Gemini).
### π AII vs. Traditional Workflows
| Task | Without AII | With AII |
|------|-------------|----------|
| **Git Workflows** | Manual commits, PRs, branches | `aii commit/pr/branch` β AI-powered git automation |
| **Translation** | Copy to browser β ChatGPT β Copy back | `aii translate "text" to spanish` β Instant result |
| **Code Generation** | Search Stack Overflow β Adapt code | `aii create python email validator` β Complete function |
| **Shell Commands** | Google syntax β Test carefully | `aii find large files` β Safe command with explanation |
| **Content Writing** | Open docs/emails separately | `aii write email declining meeting` β Professional draft |
**Result:** Stay in your terminal, maintain flow state, get instant AI assistance.
---
## β¨ Key Features
### π― Natural Language Interface
Just describe what you wantβno commands to memorize:
- **25 AI Functions**: Content generation, code operations, git workflows, shell automation, analysis, translation
- **Multi-Provider**: Anthropic Claude, OpenAI GPT, Google Gemini with automatic fallbacks
- **Smart Confirmations**: Context-aware safety checks that don't interrupt safe operations
- **Session Memory**: Persistent conversations across commands
### π‘οΈ Advanced Features
**New in v0.4.7:**
- **π Template Library**: 8 pre-built content templates for rapid generation
- Marketing: Product announcements, social media posts, launch tweets
- Development: PR descriptions, release notes
- Business: Professional emails, meeting notes, blog intros
- Variable substitution with Handlebars-style syntax
- User override system: `~/.config/aii/templates/`
- CLI: `aii template list/show/use` or natural language
- **π Usage Analytics**: Track your AI function usage with detailed statistics
- Command history analysis (7d, 30d, 90d, all time)
- Function usage breakdown with percentages
- Token and cost estimation per execution
- `aii stats` command for instant insights
- Privacy-first: Local SQLite storage, no external tracking
**New in v0.4.5:**
- **π Git PR Generator**: AI-powered pull request creation with automatic title and description
- Analyzes commit history across branches
- Generates comprehensive PR descriptions with summary, changes, and test plan
- Creates PR directly on GitHub via `gh` CLI
- **πΏ Smart Branch Naming**: Conventional branch names from natural language
- Auto-detects type: feature/, bugfix/, docs/, refactor/, etc.
- Generates clean slugs from descriptions
- Validates branch doesn't exist before creation
**New in v0.4.4:**
- **π Real-Time Response Streaming**: See LLM responses as they're generated β 60-80% faster perceived speed
- Token-by-token display with <100ms latency
- Works with Claude, GPT, and Gemini
- Markdown rendering during streaming
- Configurable with `--no-streaming` flag
- **π¨ OutputMode System**: Function-specific output formatting β Get exactly the detail you need
- **CLEAN mode** (16 functions, 67% default): Just the result (perfect for quick queries and pipeable output)
- **STANDARD mode** (5 functions, 21%): Result + metrics + simple summary (no expensive LLM analysis)
- **THINKING mode** (3 functions, 12%): Full reasoning display with LLM semantic analysis
- Smart defaults: Most functions β CLEAN (shell, translate, content), Analysis β THINKING (research, code_review)
- CLI overrides: `--clean`, `--standard`, `--thinking` or per-function config in config.yaml
- Session header only in THINKING mode (STANDARD/CLEAN are cleaner and faster)
**New in v0.4.3:**
- **β‘ Shell Autocomplete**: Tab completion for bash/zsh/fish β Type 40-60% faster
- **π Command History**: Arrow key navigation in interactive mode with 1000-command persistence
**New in v0.4.2:**
- **π Interactive Setup Wizard**: 2-minute guided setup with arrow key navigation
- **π° Real-Time Cost Tracking**: Transparent pricing for 27 AI models across 3 providers
- **βοΈ Quick Config**: Fast provider/model switching with `aii config provider/model`
**Core Features:**
- **LLM-First Architecture**: Advanced natural language understanding
- **Thinking Mode Display**: See AI reasoning, confidence scores, and token usage
- **Web Search Integration**: Research with real-time web results (DuckDuckGo, Brave, Google)
- **Health Diagnostics**: `aii doctor` command for troubleshooting setup issues
- **Debug Mode**: Complete diagnostics with 56 security patterns
---
## π Quick Start
### Installation
```bash
# With uv (recommended)
uv tool install aiiware-cli
# Or with pip
pip install aiiware-cli
# Verify
aii --version # β aiiware-cli 0.4.7
```
### Setup *(Interactive Wizard in v0.4.2)*
```bash
# Interactive setup wizard (2 minutes)
aii config init
# Verify setup
aii doctor # β Runs health checks for configuration, LLM provider, storage, etc.
```
**Features:**
- **Arrow key navigation** β Visual menu with provider descriptions
- **Automatic validation** β Tests API key before saving
- **Browser integration** β Opens API key signup pages
- **Guided experience** β Provider selection β API key β Model choice
**Troubleshooting:** If you encounter issues, run `aii doctor` to diagnose configuration problems.
**Get API Keys:**
- **Claude**: [console.anthropic.com](https://console.anthropic.com/)
- **GPT**: [platform.openai.com](https://platform.openai.com/)
- **Gemini**: [aistudio.google.com/apikey](https://aistudio.google.com/apikey)
### Shell Autocomplete *(New in v0.4.3)*
Enable tab completion for commands, flags, and parameters in bash, zsh, and fish:
```bash
# Install (auto-detects your shell)
aii install-completion
# Or specify shell explicitly
aii install-completion --shell bash
aii install-completion --shell zsh
aii install-completion --shell fish
```
**For zsh users:** Add to `~/.zshrc` if not already present:
```bash
fpath=(~/.zsh/completion $fpath)
autoload -Uz compinit && compinit
```
**Activate:**
```bash
source ~/.bashrc # bash
exec zsh # zsh (or restart terminal)
# fish auto-loads automatically
```
**Usage:**
```bash
aii tr<TAB> β aii translate
aii translate --<TAB> β --to --from --format --help
aii translate --to <TAB> β spanish french german chinese # zsh/fish
```
**Uninstall:**
```bash
aii uninstall-completion
```
### Interactive Mode with Command History *(New in v0.4.3)*
Interactive mode now includes persistent command history with arrow key navigation:
```bash
# Start interactive mode
aii
# Your commands are saved and can be recalled
> translate hello to French
Bonjour
> explain how LLMs work
[explanation...]
> β # UP arrow β recalls previous commands
> β # DOWN arrow β navigate forward
```
**Features:**
- **Arrow keys**: UP/DOWN for history, LEFT/RIGHT for editing
- **Keyboard shortcuts**: Home/End, Ctrl+A/E/K/U
- **Persistent**: 1000 commands saved at `~/.config/aii/.aii_history`
- **Zero config**: Works automatically
---
## π‘ What You Can Do
### π§ Core Use Cases
#### π **AI Git Workflows** β Complete git automation
```bash
# Smart commit messages
git add . && aii commit
# β Analyzes changes β Generates conventional commit β Shows confidence score
# Pull request creation (New in v0.4.5)
aii pr
# β Analyzes commits β Generates PR title/description β Creates on GitHub
# Smart branch naming (New in v0.4.5)
aii branch "add user authentication"
# β feature/add-user-authentication
```
#### π» **Code Generation** β Language-agnostic with best practices
```bash
aii create a python function to validate emails
# Complete with typing, docstrings, error handling, and examples
```
#### π **Translation** β Context-aware with cultural notes
```bash
# Quick answer (CLEAN mode default)
aii translate "running late for meeting" to spanish
# β "Voy a llegar tarde a la reuniΓ³n"
# With full reasoning (override with --thinking)
aii translate "running late for meeting" to spanish --thinking
# β Shows translation logic, cultural context, confidence score
```
#### π **Content Creation** β Professional writing assistance
```bash
aii write a tweet about AI breakthroughs in 2025
# β Engaging tweet with hashtags, character count validation
```
#### π **Template Library** β Rapid content generation **(New in v0.4.7)**
```bash
# List available templates
aii template list
# β 8 templates: product-announcement, tweet-launch, email-professional...
# View template details
aii template show product-announcement
# β Shows variables: product, version, key_features, etc.
# Generate from template
aii template use product-announcement \
--product "AII CLI" \
--version "v0.4.7" \
--key-features "Template library, usage analytics"
# β Professional announcement with variable substitution
# Or use natural language
aii "generate a product announcement for AII v0.4.7"
# β Detects intent and uses appropriate template
```
#### π **Usage Analytics** β Track your AI function usage **(New in v0.4.7)**
```bash
# View usage statistics
aii stats
# β Last 30 days: Total executions, top functions, usage percentages
# Weekly breakdown
aii stats --period 7d
# β Last 7 days statistics
# Function-specific analysis
aii stats --breakdown functions
# β Detailed function usage with tokens and cost estimates
# Exclude stats function from results (avoid observer effect)
aii stats --exclude-stats
# β Shows only "real work" functions, not stats checks
```
#### π **Shell Automation** β Safe command generation with explanations
```bash
aii find all python files modified this week
# β Generates command, explains safety, requests confirmation
```
#### π§ **Explanations** β Clear technical concepts
```bash
aii explain how transformers work in LLMs
# β Structured explanation with examples and analogies
```
#### π **Web Research** β Real-time information with sources
```bash
aii research latest developments in quantum computing
# β Searches web (DuckDuckGo/Brave/Google)
# β Synthesizes findings with [Source N] citations
```
#### π¨ **Output Modes** β Control detail level **(New in v0.4.4)**
```bash
# CLEAN mode: Just the answer (default for most functions)
aii translate hello to french
# β "bonjour"
aii find largest file in Downloads --clean
# Generated command: `find ~/Downloads -type f -exec du -h {} + | sort -rh | head -n 1`
# Execute? [y/N]: y
# β 1.2G bigfile.zip
# STANDARD mode: Answer + metrics (no expensive LLM analysis)
aii translate hello to french --standard
# β "bonjour"
# β β translate: (1.8s) β’ 145β 28β (173 tokens)
# β π Session Summary: β‘ Total: 3.2s β’ π° $0.000043
# THINKING mode: Full reasoning (for when you need to understand why)
aii commit --thinking
# β π aii β’ Claude β’ Functions: 23 loaded
# β [Diff analysis + reasoning + commit message + confidence + LLM insights]
```
---
## π Usage Guide
### π― Natural Language Commands
No need to memorize syntaxβjust describe what you want:
| Command Type | Example | What Happens |
|-------------|---------|--------------|
| **Translation** | `aii translate hello to spanish` | Context-aware translation with cultural notes |
| **Git** | `aii commit` / `aii pr` / `aii branch "feature"` | AI-powered commits, PRs, and branch naming |
| **Code** | `aii write a python email validator` | Complete function with types, docs, examples |
| **Content** | `aii write a tweet about AI` | Character-limited, engaging, with hashtags |
| **Shell** | `aii find large files` | Generates safe command with explanation |
| **Analysis** | `aii explain machine learning` | Structured explanation with examples |
### π§ All 25 AI Functions
<details>
<summary><b>π Content Generation</b> (5 functions) β Click to expand</summary>
- **Universal Content**: Flexible generation with tone/style customization
- **Twitter/Social**: Platform-specific posts with character limits
- **Email Templates**: Professional business communication
- **Blog/Article**: Long-form content creation
</details>
<details>
<summary><b>π» Code Operations</b> (2 functions)</summary>
- **Code Generation**: Multi-language with best practices, typing, documentation
- **Code Review**: Security analysis, style suggestions, optimization tips
</details>
<details>
<summary><b>π Git Integration</b> (5 functions) β Enhanced in v0.4.5</summary>
- **AI Commit**: Diff analysis β Conventional commit messages with thinking mode
- **PR Generator** *(New)*: Analyze commits β Create GitHub pull requests with AI descriptions
- **Smart Branch** *(New)*: Natural language β Conventional branch names (feature/, bugfix/, etc.)
- **Git Status**: Repository insights with actionable suggestions
- **Git Diff**: Intelligent change analysis and explanations
</details>
<details>
<summary><b>π Shell Operations</b> (4 functions)</summary>
- **Command Generation**: Safe shell command creation with confirmation
- **Enhanced Shell**: AI-powered execution with tool calling
- **Streaming Shell**: Real-time command output
- **Contextual Shell**: Environment-aware operations
</details>
<details>
<summary><b>π§ Analysis & Research</b> (3 functions)</summary>
- **Explain**: Technical concepts with examples and analogies
- **Summarize**: Intelligent content condensation
- **Research**: Web-based research with source validation
</details>
<details>
<summary><b>π Translation & Language</b> (2 functions)</summary>
- **Translation**: Multi-language with cultural context
- **Language Detection**: Automatic language identification
</details>
<details>
<summary><b>π Context Management</b> (3 functions)</summary>
- **Git Context**: Repository history and structure analysis
- **File Context**: Project structure and file system insights
- **System Context**: Environment information gathering
</details>
<details>
<summary><b>βοΈ System</b> (2 functions)</summary>
- **Help**: Dynamic assistance based on your context
- **Clarification**: Handles ambiguous requests intelligently
</details>
### ποΈ Advanced Features
#### **OutputMode System** *(New in v0.4.4)*
Control output detail with function-specific defaults and CLI overrides:
- **CLEAN mode** (default for translate, explain, summarize): Just the result
- **STANDARD mode**: Result + metrics + session summary
- **THINKING mode** (default for git commit, code review): Full AI reasoning
```bash
# Defaults work intelligently
aii translate hello to spanish # β "hola" (CLEAN default)
aii commit # β Full reasoning (THINKING default)
# Override any function
aii translate hello to spanish --thinking # β See reasoning
aii commit --clean # β Quick commit
# CLI flags: --clean, --standard, --thinking, --verbose
```
#### **Smart Risk-Based Confirmations**
- **Safe operations** (translate, explain): Execute immediately, no interruption
- **Risky operations** (shell, git): Show command preview + safety notes + confirmation
#### **Token Tracking & Cost Transparency** *(Enhanced in v0.4.2)*
Monitor LLM usage and costs in real-time across all operations:
```bash
# π Session: 3.2s β’ Tokens: 142β 28β (170 total) β’ π° $0.000049
# Shows: execution time, input/output tokens, exact cost
```
#### **Session Persistence**
```bash
aii explain quantum computing # Auto-saves as session_abc123
aii --continue give me more examples # Resumes conversation
```
#### **Debug Mode**
```bash
AII_DEBUG=1 aii commit # Shows: security patterns, LLM recognition, token breakdown
```
---
## π¨ Real-World Examples
### **AI Git Commit**
```bash
git add . && aii commit
# π§ Thinking: Analyzing changes... console.log added for debugging
# π» Commit: feat(app): add console logging for debugging
# π― Confidence: 89% β’ Tokens: 245β 67β
# Proceed? (y/n): y
# β
Committed: a1b2c3d
```
### **Context-Aware Translation**
```bash
aii translate "running late for meeting" to spanish
# π§ Converting informal business expression, maintaining urgency...
# π "Voy a llegar tarde a la reuniΓ³n"
# π Informal but professional. Alternative: "Me voy a retrasar" (more formal)
# π― Confidence: 94%
```
### **Code Generation with Best Practices**
```bash
aii create a python email validator with error handling
# π» Returns complete function with:
# - RFC 5322 regex pattern
# - Type hints (typing.bool)
# - Comprehensive docstring
# - Error handling (TypeError, re.error)
# - Example usage with test cases
# π― Confidence: 92% β’ 445 tokens generated
```
---
## βοΈ Configuration
```bash
# Interactive setup
aii config init # Guided provider + API key setup (v0.4.2: arrow key menu)
aii config show # View current config
aii config validate # Check configuration
# Quick config commands (New in v0.4.2)
aii config provider # Switch LLM provider (Claude/GPT/Gemini)
aii config model # Change model within provider
aii config web-search # Configure web search integration
# Command-line options
aii --help # Show help
aii --version # Version info
aii --continue "msg" # Resume last session
```
---
## π Troubleshooting
**Quick Diagnosis:** Run `aii doctor` to check your setup automatically.
| Issue | Solution |
|-------|----------|
| β `ANTHROPIC_API_KEY not found` | Run `aii config init` or `export ANTHROPIC_API_KEY="key"` |
| β `aii command not found` | Restart terminal or check PATH: `uv tool install aiiware-cli` |
| β οΈ Slow responses | Check network or try different provider: `aii config init` |
| π Need diagnostics | Run `aii doctor` or enable debug: `AII_DEBUG=1 aii your-command` |
| π LLM timeout errors | Retry logic automatic in v0.4.1, check `aii doctor` for connectivity |
| β οΈ Web search not working | Configure Brave Search API key (optional, free DuckDuckGo fallback available) |
---
## π§ Architecture
AII uses LLM-first design with:
- **Intent Recognition**: Natural language β function mapping via LLMs
- **Function Plugins**: 25 extensible functions across 8 categories
- **Smart Risk Assessment**: Context-aware confirmations
- **Multi-Provider**: Anthropic Claude, OpenAI GPT, Google Gemini
- **Session Persistence**: Conversation memory and metrics
---
## π Links & Resources
- **π¦ [PyPI Package](https://pypi.org/project/aiiware-cli)** β Official distribution
- **π Documentation** β Comprehensive examples in this README
- **π Releases** β Version history on PyPI
---
## π License & Credits
**License:** Apache 2.0 β’ Copyright 2025-present aiiware.com
**Built with:**
- [Pydantic AI](https://ai.pydantic.dev/) β Reliable LLM integration framework
- [Anthropic Claude](https://www.anthropic.com/claude) β’ [OpenAI GPT](https://openai.com/) β’ [Google Gemini](https://ai.google.dev/)
---
<div align="center">
**Made with β€οΈ for developers who want AI-powered assistance without leaving the terminal**
[β¬ Back to top](#aii--ai-powered-terminal-assistant-for-developers)
</div>
Raw data
{
"_id": null,
"home_page": null,
"name": "aiiware-cli",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.10",
"maintainer_email": "aii Dev <aii-dev@aiiware.com>",
"keywords": "ai, ai-assistant, ai-cli, ai-code-review, ai-commands, ai-git-commit, ai-translation, ai-writing, anthropic, automation, chatgpt, chatgpt-cli, claude, cli, code-analysis, code-assistant, code-explainer, code-generation, coding-assistant, command-line, commit-message-generator, content-generation, conversational-ai, cross-platform, developer-tools, devtools, gemini, git-assistant, git-commit, google-ai, gpt, llm, llm-cli, natural-language, openai, productivity, pydantic-ai, python, python-cli, shell-automation, terminal, terminal-assistant, translation, translator, workflow-automation, writing-assistant",
"author": null,
"author_email": "aii Dev <aii-dev@aiiware.com>",
"download_url": "https://files.pythonhosted.org/packages/98/1e/a4d8da6f57b6974ad70e8aeb10b266fb7f5113e1a6cd8afe473444ffe6aa/aiiware_cli-0.4.7.tar.gz",
"platform": null,
"description": "# AII \u2014 AI-Powered Terminal Assistant\n\n**AII** (pronounced *\"eye\"*) is your **intelligent command-line companion** that brings Claude, GPT, and Gemini directly into your terminal through natural language commands.\n\n**\ud83c\udf89 Version 0.4.7** \u2014 Template Library & Analytics | **v0.4.5** \u2014 Git PR & Branch | **v0.4.4** \u2014 Streaming + Output modes\n\n## Why AII?\n\n**One CLI. 25 AI Functions. Natural Language. Zero Configuration Headaches.**\n\nStop context-switching between your terminal and ChatGPT. Get instant AI assistance for git commits, code generation, translation, content writing, shell automation, and analysis\u2014all without leaving your terminal.\n\n**Perfect for:**\n- **Developers**: AI-powered git commits, code generation, and shell automation\n- **Creators**: Content writing, translation, and social media posts\n- **Professionals**: Email templates, document analysis, and research assistance\n\nBuilt with the Pydantic AI framework for reliable, multi-provider LLM integration (Anthropic Claude, OpenAI GPT, Google Gemini).\n\n### \ud83d\udcca AII vs. Traditional Workflows\n\n| Task | Without AII | With AII |\n|------|-------------|----------|\n| **Git Workflows** | Manual commits, PRs, branches | `aii commit/pr/branch` \u2192 AI-powered git automation |\n| **Translation** | Copy to browser \u2192 ChatGPT \u2192 Copy back | `aii translate \"text\" to spanish` \u2192 Instant result |\n| **Code Generation** | Search Stack Overflow \u2192 Adapt code | `aii create python email validator` \u2192 Complete function |\n| **Shell Commands** | Google syntax \u2192 Test carefully | `aii find large files` \u2192 Safe command with explanation |\n| **Content Writing** | Open docs/emails separately | `aii write email declining meeting` \u2192 Professional draft |\n\n**Result:** Stay in your terminal, maintain flow state, get instant AI assistance.\n\n---\n\n## \u2728 Key Features\n\n### \ud83c\udfaf Natural Language Interface\nJust describe what you want\u2014no commands to memorize:\n- **25 AI Functions**: Content generation, code operations, git workflows, shell automation, analysis, translation\n- **Multi-Provider**: Anthropic Claude, OpenAI GPT, Google Gemini with automatic fallbacks\n- **Smart Confirmations**: Context-aware safety checks that don't interrupt safe operations\n- **Session Memory**: Persistent conversations across commands\n\n### \ud83d\udee1\ufe0f Advanced Features\n\n**New in v0.4.7:**\n- **\ud83d\udcdd Template Library**: 8 pre-built content templates for rapid generation\n - Marketing: Product announcements, social media posts, launch tweets\n - Development: PR descriptions, release notes\n - Business: Professional emails, meeting notes, blog intros\n - Variable substitution with Handlebars-style syntax\n - User override system: `~/.config/aii/templates/`\n - CLI: `aii template list/show/use` or natural language\n- **\ud83d\udcca Usage Analytics**: Track your AI function usage with detailed statistics\n - Command history analysis (7d, 30d, 90d, all time)\n - Function usage breakdown with percentages\n - Token and cost estimation per execution\n - `aii stats` command for instant insights\n - Privacy-first: Local SQLite storage, no external tracking\n\n**New in v0.4.5:**\n- **\ud83d\udd00 Git PR Generator**: AI-powered pull request creation with automatic title and description\n - Analyzes commit history across branches\n - Generates comprehensive PR descriptions with summary, changes, and test plan\n - Creates PR directly on GitHub via `gh` CLI\n- **\ud83c\udf3f Smart Branch Naming**: Conventional branch names from natural language\n - Auto-detects type: feature/, bugfix/, docs/, refactor/, etc.\n - Generates clean slugs from descriptions\n - Validates branch doesn't exist before creation\n\n**New in v0.4.4:**\n- **\ud83c\udf0a Real-Time Response Streaming**: See LLM responses as they're generated \u2014 60-80% faster perceived speed\n - Token-by-token display with <100ms latency\n - Works with Claude, GPT, and Gemini\n - Markdown rendering during streaming\n - Configurable with `--no-streaming` flag\n- **\ud83c\udfa8 OutputMode System**: Function-specific output formatting \u2014 Get exactly the detail you need\n - **CLEAN mode** (16 functions, 67% default): Just the result (perfect for quick queries and pipeable output)\n - **STANDARD mode** (5 functions, 21%): Result + metrics + simple summary (no expensive LLM analysis)\n - **THINKING mode** (3 functions, 12%): Full reasoning display with LLM semantic analysis\n - Smart defaults: Most functions \u2192 CLEAN (shell, translate, content), Analysis \u2192 THINKING (research, code_review)\n - CLI overrides: `--clean`, `--standard`, `--thinking` or per-function config in config.yaml\n - Session header only in THINKING mode (STANDARD/CLEAN are cleaner and faster)\n\n**New in v0.4.3:**\n- **\u26a1 Shell Autocomplete**: Tab completion for bash/zsh/fish \u2014 Type 40-60% faster\n- **\ud83d\udcdc Command History**: Arrow key navigation in interactive mode with 1000-command persistence\n\n**New in v0.4.2:**\n- **\ud83d\ude80 Interactive Setup Wizard**: 2-minute guided setup with arrow key navigation\n- **\ud83d\udcb0 Real-Time Cost Tracking**: Transparent pricing for 27 AI models across 3 providers\n- **\u2699\ufe0f Quick Config**: Fast provider/model switching with `aii config provider/model`\n\n**Core Features:**\n- **LLM-First Architecture**: Advanced natural language understanding\n- **Thinking Mode Display**: See AI reasoning, confidence scores, and token usage\n- **Web Search Integration**: Research with real-time web results (DuckDuckGo, Brave, Google)\n- **Health Diagnostics**: `aii doctor` command for troubleshooting setup issues\n- **Debug Mode**: Complete diagnostics with 56 security patterns\n\n---\n\n## \ud83d\ude80 Quick Start\n\n### Installation\n\n```bash\n# With uv (recommended)\nuv tool install aiiware-cli\n\n# Or with pip\npip install aiiware-cli\n\n# Verify\naii --version # \u2192 aiiware-cli 0.4.7\n```\n\n### Setup *(Interactive Wizard in v0.4.2)*\n\n```bash\n# Interactive setup wizard (2 minutes)\naii config init\n\n# Verify setup\naii doctor # \u2192 Runs health checks for configuration, LLM provider, storage, etc.\n```\n\n**Features:**\n- **Arrow key navigation** \u2014 Visual menu with provider descriptions\n- **Automatic validation** \u2014 Tests API key before saving\n- **Browser integration** \u2014 Opens API key signup pages\n- **Guided experience** \u2014 Provider selection \u2192 API key \u2192 Model choice\n\n**Troubleshooting:** If you encounter issues, run `aii doctor` to diagnose configuration problems.\n\n**Get API Keys:**\n- **Claude**: [console.anthropic.com](https://console.anthropic.com/)\n- **GPT**: [platform.openai.com](https://platform.openai.com/)\n- **Gemini**: [aistudio.google.com/apikey](https://aistudio.google.com/apikey)\n\n### Shell Autocomplete *(New in v0.4.3)*\n\nEnable tab completion for commands, flags, and parameters in bash, zsh, and fish:\n\n```bash\n# Install (auto-detects your shell)\naii install-completion\n\n# Or specify shell explicitly\naii install-completion --shell bash\naii install-completion --shell zsh\naii install-completion --shell fish\n```\n\n**For zsh users:** Add to `~/.zshrc` if not already present:\n```bash\nfpath=(~/.zsh/completion $fpath)\nautoload -Uz compinit && compinit\n```\n\n**Activate:**\n```bash\nsource ~/.bashrc # bash\nexec zsh # zsh (or restart terminal)\n# fish auto-loads automatically\n```\n\n**Usage:**\n```bash\naii tr<TAB> \u2192 aii translate\naii translate --<TAB> \u2192 --to --from --format --help\naii translate --to <TAB> \u2192 spanish french german chinese # zsh/fish\n```\n\n**Uninstall:**\n```bash\naii uninstall-completion\n```\n\n### Interactive Mode with Command History *(New in v0.4.3)*\n\nInteractive mode now includes persistent command history with arrow key navigation:\n\n```bash\n# Start interactive mode\naii\n\n# Your commands are saved and can be recalled\n> translate hello to French\nBonjour\n\n> explain how LLMs work\n[explanation...]\n\n> \u2191 # UP arrow \u2014 recalls previous commands\n> \u2193 # DOWN arrow \u2014 navigate forward\n```\n\n**Features:**\n- **Arrow keys**: UP/DOWN for history, LEFT/RIGHT for editing\n- **Keyboard shortcuts**: Home/End, Ctrl+A/E/K/U\n- **Persistent**: 1000 commands saved at `~/.config/aii/.aii_history`\n- **Zero config**: Works automatically\n\n---\n\n## \ud83d\udca1 What You Can Do\n\n### \ud83d\udd27 Core Use Cases\n\n#### \ud83d\udd00 **AI Git Workflows** \u2014 Complete git automation\n```bash\n# Smart commit messages\ngit add . && aii commit\n# \u2192 Analyzes changes \u2192 Generates conventional commit \u2192 Shows confidence score\n\n# Pull request creation (New in v0.4.5)\naii pr\n# \u2192 Analyzes commits \u2192 Generates PR title/description \u2192 Creates on GitHub\n\n# Smart branch naming (New in v0.4.5)\naii branch \"add user authentication\"\n# \u2192 feature/add-user-authentication\n```\n\n#### \ud83d\udcbb **Code Generation** \u2014 Language-agnostic with best practices\n```bash\naii create a python function to validate emails\n# Complete with typing, docstrings, error handling, and examples\n```\n\n#### \ud83c\udf0d **Translation** \u2014 Context-aware with cultural notes\n```bash\n# Quick answer (CLEAN mode default)\naii translate \"running late for meeting\" to spanish\n# \u2192 \"Voy a llegar tarde a la reuni\u00f3n\"\n\n# With full reasoning (override with --thinking)\naii translate \"running late for meeting\" to spanish --thinking\n# \u2192 Shows translation logic, cultural context, confidence score\n```\n\n#### \ud83d\udcdd **Content Creation** \u2014 Professional writing assistance\n```bash\naii write a tweet about AI breakthroughs in 2025\n# \u2192 Engaging tweet with hashtags, character count validation\n```\n\n#### \ud83d\udcda **Template Library** \u2014 Rapid content generation **(New in v0.4.7)**\n```bash\n# List available templates\naii template list\n# \u2192 8 templates: product-announcement, tweet-launch, email-professional...\n\n# View template details\naii template show product-announcement\n# \u2192 Shows variables: product, version, key_features, etc.\n\n# Generate from template\naii template use product-announcement \\\n --product \"AII CLI\" \\\n --version \"v0.4.7\" \\\n --key-features \"Template library, usage analytics\"\n# \u2192 Professional announcement with variable substitution\n\n# Or use natural language\naii \"generate a product announcement for AII v0.4.7\"\n# \u2192 Detects intent and uses appropriate template\n```\n\n#### \ud83d\udcca **Usage Analytics** \u2014 Track your AI function usage **(New in v0.4.7)**\n```bash\n# View usage statistics\naii stats\n# \u2192 Last 30 days: Total executions, top functions, usage percentages\n\n# Weekly breakdown\naii stats --period 7d\n# \u2192 Last 7 days statistics\n\n# Function-specific analysis\naii stats --breakdown functions\n# \u2192 Detailed function usage with tokens and cost estimates\n\n# Exclude stats function from results (avoid observer effect)\naii stats --exclude-stats\n# \u2192 Shows only \"real work\" functions, not stats checks\n```\n\n#### \ud83d\udc1a **Shell Automation** \u2014 Safe command generation with explanations\n```bash\naii find all python files modified this week\n# \u2192 Generates command, explains safety, requests confirmation\n```\n\n#### \ud83e\udde0 **Explanations** \u2014 Clear technical concepts\n```bash\naii explain how transformers work in LLMs\n# \u2192 Structured explanation with examples and analogies\n```\n\n#### \ud83d\udd0d **Web Research** \u2014 Real-time information with sources\n```bash\naii research latest developments in quantum computing\n# \u2192 Searches web (DuckDuckGo/Brave/Google)\n# \u2192 Synthesizes findings with [Source N] citations\n```\n\n#### \ud83c\udfa8 **Output Modes** \u2014 Control detail level **(New in v0.4.4)**\n```bash\n# CLEAN mode: Just the answer (default for most functions)\naii translate hello to french\n# \u2192 \"bonjour\"\n\naii find largest file in Downloads --clean\n# Generated command: `find ~/Downloads -type f -exec du -h {} + | sort -rh | head -n 1`\n# Execute? [y/N]: y\n# \u2192 1.2G bigfile.zip\n\n# STANDARD mode: Answer + metrics (no expensive LLM analysis)\naii translate hello to french --standard\n# \u2192 \"bonjour\"\n# \u2192 \u2713 translate: (1.8s) \u2022 145\u2197 28\u2198 (173 tokens)\n# \u2192 \ud83d\udcca Session Summary: \u26a1 Total: 3.2s \u2022 \ud83d\udcb0 $0.000043\n\n# THINKING mode: Full reasoning (for when you need to understand why)\naii commit --thinking\n# \u2192 \ud83d\udd0d aii \u2022 Claude \u2022 Functions: 23 loaded\n# \u2192 [Diff analysis + reasoning + commit message + confidence + LLM insights]\n```\n\n---\n\n## \ud83d\udcd6 Usage Guide\n\n### \ud83c\udfaf Natural Language Commands\n\nNo need to memorize syntax\u2014just describe what you want:\n\n| Command Type | Example | What Happens |\n|-------------|---------|--------------|\n| **Translation** | `aii translate hello to spanish` | Context-aware translation with cultural notes |\n| **Git** | `aii commit` / `aii pr` / `aii branch \"feature\"` | AI-powered commits, PRs, and branch naming |\n| **Code** | `aii write a python email validator` | Complete function with types, docs, examples |\n| **Content** | `aii write a tweet about AI` | Character-limited, engaging, with hashtags |\n| **Shell** | `aii find large files` | Generates safe command with explanation |\n| **Analysis** | `aii explain machine learning` | Structured explanation with examples |\n\n### \ud83d\udd27 All 25 AI Functions\n\n<details>\n<summary><b>\ud83d\udcdd Content Generation</b> (5 functions) \u2014 Click to expand</summary>\n\n- **Universal Content**: Flexible generation with tone/style customization\n- **Twitter/Social**: Platform-specific posts with character limits\n- **Email Templates**: Professional business communication\n- **Blog/Article**: Long-form content creation\n</details>\n\n<details>\n<summary><b>\ud83d\udcbb Code Operations</b> (2 functions)</summary>\n\n- **Code Generation**: Multi-language with best practices, typing, documentation\n- **Code Review**: Security analysis, style suggestions, optimization tips\n</details>\n\n<details>\n<summary><b>\ud83d\udd00 Git Integration</b> (5 functions) \u2014 Enhanced in v0.4.5</summary>\n\n- **AI Commit**: Diff analysis \u2192 Conventional commit messages with thinking mode\n- **PR Generator** *(New)*: Analyze commits \u2192 Create GitHub pull requests with AI descriptions\n- **Smart Branch** *(New)*: Natural language \u2192 Conventional branch names (feature/, bugfix/, etc.)\n- **Git Status**: Repository insights with actionable suggestions\n- **Git Diff**: Intelligent change analysis and explanations\n</details>\n\n<details>\n<summary><b>\ud83d\udc1a Shell Operations</b> (4 functions)</summary>\n\n- **Command Generation**: Safe shell command creation with confirmation\n- **Enhanced Shell**: AI-powered execution with tool calling\n- **Streaming Shell**: Real-time command output\n- **Contextual Shell**: Environment-aware operations\n</details>\n\n<details>\n<summary><b>\ud83e\udde0 Analysis & Research</b> (3 functions)</summary>\n\n- **Explain**: Technical concepts with examples and analogies\n- **Summarize**: Intelligent content condensation\n- **Research**: Web-based research with source validation\n</details>\n\n<details>\n<summary><b>\ud83c\udf0d Translation & Language</b> (2 functions)</summary>\n\n- **Translation**: Multi-language with cultural context\n- **Language Detection**: Automatic language identification\n</details>\n\n<details>\n<summary><b>\ud83d\udcca Context Management</b> (3 functions)</summary>\n\n- **Git Context**: Repository history and structure analysis\n- **File Context**: Project structure and file system insights\n- **System Context**: Environment information gathering\n</details>\n\n<details>\n<summary><b>\u2699\ufe0f System</b> (2 functions)</summary>\n\n- **Help**: Dynamic assistance based on your context\n- **Clarification**: Handles ambiguous requests intelligently\n</details>\n\n### \ud83c\udf9b\ufe0f Advanced Features\n\n#### **OutputMode System** *(New in v0.4.4)*\nControl output detail with function-specific defaults and CLI overrides:\n\n- **CLEAN mode** (default for translate, explain, summarize): Just the result\n- **STANDARD mode**: Result + metrics + session summary\n- **THINKING mode** (default for git commit, code review): Full AI reasoning\n\n```bash\n# Defaults work intelligently\naii translate hello to spanish # \u2192 \"hola\" (CLEAN default)\naii commit # \u2192 Full reasoning (THINKING default)\n\n# Override any function\naii translate hello to spanish --thinking # \u2192 See reasoning\naii commit --clean # \u2192 Quick commit\n\n# CLI flags: --clean, --standard, --thinking, --verbose\n```\n\n#### **Smart Risk-Based Confirmations**\n- **Safe operations** (translate, explain): Execute immediately, no interruption\n- **Risky operations** (shell, git): Show command preview + safety notes + confirmation\n\n#### **Token Tracking & Cost Transparency** *(Enhanced in v0.4.2)*\nMonitor LLM usage and costs in real-time across all operations:\n```bash\n# \ud83d\udcca Session: 3.2s \u2022 Tokens: 142\u2197 28\u2198 (170 total) \u2022 \ud83d\udcb0 $0.000049\n# Shows: execution time, input/output tokens, exact cost\n```\n\n#### **Session Persistence**\n```bash\naii explain quantum computing # Auto-saves as session_abc123\naii --continue give me more examples # Resumes conversation\n```\n\n#### **Debug Mode**\n```bash\nAII_DEBUG=1 aii commit # Shows: security patterns, LLM recognition, token breakdown\n```\n\n---\n\n## \ud83c\udfa8 Real-World Examples\n\n### **AI Git Commit**\n\n```bash\ngit add . && aii commit\n\n# \ud83e\udde0 Thinking: Analyzing changes... console.log added for debugging\n# \ud83d\udcbb Commit: feat(app): add console logging for debugging\n# \ud83c\udfaf Confidence: 89% \u2022 Tokens: 245\u2197 67\u2198\n# Proceed? (y/n): y\n# \u2705 Committed: a1b2c3d\n```\n\n### **Context-Aware Translation**\n\n```bash\naii translate \"running late for meeting\" to spanish\n\n# \ud83e\udde0 Converting informal business expression, maintaining urgency...\n# \ud83c\udf10 \"Voy a llegar tarde a la reuni\u00f3n\"\n# \ud83d\udcdd Informal but professional. Alternative: \"Me voy a retrasar\" (more formal)\n# \ud83c\udfaf Confidence: 94%\n```\n\n### **Code Generation with Best Practices**\n\n```bash\naii create a python email validator with error handling\n\n# \ud83d\udcbb Returns complete function with:\n# - RFC 5322 regex pattern\n# - Type hints (typing.bool)\n# - Comprehensive docstring\n# - Error handling (TypeError, re.error)\n# - Example usage with test cases\n# \ud83c\udfaf Confidence: 92% \u2022 445 tokens generated\n```\n\n---\n\n## \u2699\ufe0f Configuration\n\n```bash\n# Interactive setup\naii config init # Guided provider + API key setup (v0.4.2: arrow key menu)\naii config show # View current config\naii config validate # Check configuration\n\n# Quick config commands (New in v0.4.2)\naii config provider # Switch LLM provider (Claude/GPT/Gemini)\naii config model # Change model within provider\naii config web-search # Configure web search integration\n\n# Command-line options\naii --help # Show help\naii --version # Version info\naii --continue \"msg\" # Resume last session\n```\n\n---\n\n## \ud83d\udc1b Troubleshooting\n\n**Quick Diagnosis:** Run `aii doctor` to check your setup automatically.\n\n| Issue | Solution |\n|-------|----------|\n| \u274c `ANTHROPIC_API_KEY not found` | Run `aii config init` or `export ANTHROPIC_API_KEY=\"key\"` |\n| \u274c `aii command not found` | Restart terminal or check PATH: `uv tool install aiiware-cli` |\n| \u26a0\ufe0f Slow responses | Check network or try different provider: `aii config init` |\n| \ud83d\udd0d Need diagnostics | Run `aii doctor` or enable debug: `AII_DEBUG=1 aii your-command` |\n| \ud83d\udd04 LLM timeout errors | Retry logic automatic in v0.4.1, check `aii doctor` for connectivity |\n| \u26a0\ufe0f Web search not working | Configure Brave Search API key (optional, free DuckDuckGo fallback available) |\n\n---\n\n## \ud83d\udd27 Architecture\n\nAII uses LLM-first design with:\n- **Intent Recognition**: Natural language \u2192 function mapping via LLMs\n- **Function Plugins**: 25 extensible functions across 8 categories\n- **Smart Risk Assessment**: Context-aware confirmations\n- **Multi-Provider**: Anthropic Claude, OpenAI GPT, Google Gemini\n- **Session Persistence**: Conversation memory and metrics\n\n---\n\n## \ud83d\udcda Links & Resources\n\n- **\ud83d\udce6 [PyPI Package](https://pypi.org/project/aiiware-cli)** \u2014 Official distribution\n- **\ud83d\udcd6 Documentation** \u2014 Comprehensive examples in this README\n- **\ud83d\udd04 Releases** \u2014 Version history on PyPI\n\n---\n\n## \ud83d\udcc4 License & Credits\n\n**License:** Apache 2.0 \u2022 Copyright 2025-present aiiware.com\n\n**Built with:**\n- [Pydantic AI](https://ai.pydantic.dev/) \u2014 Reliable LLM integration framework\n- [Anthropic Claude](https://www.anthropic.com/claude) \u2022 [OpenAI GPT](https://openai.com/) \u2022 [Google Gemini](https://ai.google.dev/)\n\n---\n\n<div align=\"center\">\n\n**Made with \u2764\ufe0f for developers who want AI-powered assistance without leaving the terminal**\n\n[\u2b06 Back to top](#aii--ai-powered-terminal-assistant-for-developers)\n\n</div>\n",
"bugtrack_url": null,
"license": "Apache-2.0",
"summary": "AI-powered terminal assistant for developers: natural language commands for git commits, code generation, translation, content writing, shell automation, and analysis. Supports Claude, GPT, and Gemini. Now with shell autocomplete!",
"version": "0.4.7",
"project_urls": null,
"split_keywords": [
"ai",
" ai-assistant",
" ai-cli",
" ai-code-review",
" ai-commands",
" ai-git-commit",
" ai-translation",
" ai-writing",
" anthropic",
" automation",
" chatgpt",
" chatgpt-cli",
" claude",
" cli",
" code-analysis",
" code-assistant",
" code-explainer",
" code-generation",
" coding-assistant",
" command-line",
" commit-message-generator",
" content-generation",
" conversational-ai",
" cross-platform",
" developer-tools",
" devtools",
" gemini",
" git-assistant",
" git-commit",
" google-ai",
" gpt",
" llm",
" llm-cli",
" natural-language",
" openai",
" productivity",
" pydantic-ai",
" python",
" python-cli",
" shell-automation",
" terminal",
" terminal-assistant",
" translation",
" translator",
" workflow-automation",
" writing-assistant"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "bbe72a844c0cde702ffa66ba0b7e81b00d3fec2e8f64e41e7ce5614b41788821",
"md5": "637122267718996e3e812e3327ca9df2",
"sha256": "ff6706ee47fd2d1b660025c12d21180b31db3d117beb11e2e536364f2abf8175"
},
"downloads": -1,
"filename": "aiiware_cli-0.4.7-py3-none-any.whl",
"has_sig": false,
"md5_digest": "637122267718996e3e812e3327ca9df2",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.10",
"size": 354671,
"upload_time": "2025-10-06T10:43:12",
"upload_time_iso_8601": "2025-10-06T10:43:12.783429Z",
"url": "https://files.pythonhosted.org/packages/bb/e7/2a844c0cde702ffa66ba0b7e81b00d3fec2e8f64e41e7ce5614b41788821/aiiware_cli-0.4.7-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "981ea4d8da6f57b6974ad70e8aeb10b266fb7f5113e1a6cd8afe473444ffe6aa",
"md5": "a21050258e4f5132f30e27adbab2f36c",
"sha256": "8ec0fba872d288c0924aa2e7b5db8996a238559e5c645333b6d008e836b07146"
},
"downloads": -1,
"filename": "aiiware_cli-0.4.7.tar.gz",
"has_sig": false,
"md5_digest": "a21050258e4f5132f30e27adbab2f36c",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.10",
"size": 294858,
"upload_time": "2025-10-06T10:43:14",
"upload_time_iso_8601": "2025-10-06T10:43:14.745898Z",
"url": "https://files.pythonhosted.org/packages/98/1e/a4d8da6f57b6974ad70e8aeb10b266fb7f5113e1a6cd8afe473444ffe6aa/aiiware_cli-0.4.7.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-10-06 10:43:14",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "aiiware-cli"
}