lmapp 0.3.1

Local LLM Made Simple - Consumer-friendly AI assistant

LMAPP

Local LLM CLI – AI everywhere. Easy, simple, and undeniable.
Online or offline. The future is yours to command.

v0.3.0-beta - Production Ready. Fully Featured. Free.

See Demo & Features for examples and use cases.

---

🚀 Quick Start

Full installation and setup: see QUICKSTART.md. Customize your AI's behavior: see Roles & Workflows Guide.

Everyday commands:

lmapp chat          # Start chatting locally
lmapp status        # Check backend/model status
lmapp config show   # View current configuration

More examples and demos: see DEMO.md.

---

🎯 Features

💬 Chat

$ lmapp chat --model mistral
╔════════════════════════════════════════════╗
║        Chat with Mistral (Local)           ║
╚════════════════════════════════════════════╝

You: Explain quantum computing in simple terms

AI: Quantum computers use quantum bits (qubits) instead of regular bits.
While regular bits are 0 or 1, qubits can be both at once (superposition).
This lets them solve certain problems exponentially faster...

You: What are the use cases?

AI: Key use cases include:
  • Drug discovery (molecular simulation)
  • Finance (portfolio optimization)
  • Cryptography (breaking encryption)
  • Machine learning (optimization)

🔍 RAG (Semantic Search)

$ lmapp rag index ~/my_docs
📁 Indexing documents...
✓ Processed: README.md (1,234 tokens)
✓ Processed: GUIDE.pdf (5,678 tokens)
✓ Processed: NOTES.txt (892 tokens)
✓ Index created: 7,804 tokens in 12 documents

$ lmapp rag search "how to optimize python code"
📊 Search Results (3 matches):

1. GUIDE.pdf - Line 45 (score: 0.92)
   "Optimization techniques include: list comprehensions,
    caching, and using built-in functions instead of loops"

2. NOTES.txt - Line 12 (score: 0.88)
   "Profile code with cProfile before optimizing"

3. README.md - Line 89 (score: 0.81)
   "Performance tips for production code"

$ lmapp chat --with-context
You: Summarize the best Python optimization tips from my docs

AI: Based on your documents, here are the key optimization tips:
  1. Use list comprehensions instead of loops
  2. Profile with cProfile before optimizing
  3. Leverage built-in functions (map, filter, etc.)
  4. Implement caching for expensive operations

📦 Batch Processing

$ lmapp batch create inputs.json
Processing 5 queries in batch...
[████████████████████] 100% (5/5)

Job created: batch_20250211_143022
Estimated time: 45 seconds

$ lmapp batch results batch_20250211_143022 --json
{
  "job_id": "batch_20250211_143022",
  "status": "completed",
  "results": [
    {"input": "Explain AI", "output": "AI is..."},
    {"input": "What is ML?", "output": "Machine learning..."},
    ...
  ],
  "completed_at": "2025-02-11T14:30:47Z"
}

🔌 Plugins

$ lmapp plugin list
Available Plugins:
  ✓ translator     - Real-time translation (8 languages)
  ✓ summarizer     - Extract key points from long text
  ✓ code-reviewer  - Analyze code and suggest improvements
  ✓ sql-generator  - Write SQL queries from descriptions
  ✓ regex-helper   - Build and test regex patterns
  ✓ json-validator - Validate and format JSON
  ✓ git-helper     - Explain git commands and operations
  ✓ api-tester     - Test REST APIs interactively

$ lmapp plugin install translator
Installing translator plugin...
✓ Downloaded (245 KB)
✓ Installed successfully
Ready to use: lmapp translate --help

$ lmapp translate --text "Hello World" --to spanish
Translation (Spanish):
"¡Hola Mundo!"

⚙️ Configuration

$ lmapp config show
Current Configuration:
  Model: mistral (7B)
  Temperature: 0.7
  Max Tokens: 2048
  Context Size: 4096
  System Prompt: You are a helpful AI assistant

$ lmapp config set temperature 0.3
✓ Configuration updated

$ lmapp config --set-prompt
Enter your custom system prompt:
> You are a Python expert. Help with code, explain concepts clearly.
✓ System prompt saved

$ lmapp status
Status Report:
  ✓ Backend: Ollama (running)
  ✓ Model: mistral (7.4B)
  ✓ Memory: 6.2 GB / 16 GB
  ✓ Performance: 45 tokens/sec

---

💡 Who Is This For?

Perfect Fit

• Developers - Code explanations, debugging, documentation, CLI workflows
• Students & Researchers - Study partner, research assistance, offline-first
• SysAdmins - Command lookups, automation scripts, system analysis
• Professionals - Writing, analysis, research, note-taking
• Privacy-Conscious Users - Want AI without cloud dependencies
• Anyone who values control over convenience

---

---

📖 Basic Usage

# Start chat
lmapp chat

# Use specific model
lmapp chat --model mistral

# Check status
lmapp status

# View configuration
lmapp config show

Supported Backends: Ollama, llamafile (auto-detected). Extensible architecture supports custom backends.

See QUICKSTART.md for complete usage guide.

---

✅ Quality & Features

• 🧪 587 tests (100% coverage)
• 🔒 100% private (no cloud, no tracking)
• ⚡ Fast & lightweight (<200ms startup)
• 🔌 8 production plugins
• 🔍 RAG system (semantic search)
• 📦 Batch processing
• 💾 Session persistence
• 🌐 Web UI (optional)

---

🔐 Privacy & Security

• 100% Local - Everything runs on your device
• No Cloud - No internet after setup
• No Telemetry - Zero tracking
• Open Source - MIT licensed
• Your Data - You own it all

---

🗺️ Roadmap

v0.3.0 (Current) - Production ready
v0.4.0+ - Mobile/desktop apps, team features, enterprise tier

---

🤝 Contributing

Help wanted! See Contributing Guide for code contributions, bug reports, or feature ideas.

---

All contributions welcome: bug fixes, features, documentation, tests, and ideas.

---

💬 Support

• Found a bug? Open an Issue
• Questions? See Troubleshooting Guide
• Discussions? Use GitHub Discussions

---

⚙️ Troubleshooting

Issue	Solution
command not found	Add ~/.local/bin to $PATH or use pipx install lmapp
ModuleNotFoundError	Reinstall: pip install --upgrade lmapp
Debian/Ubuntu issues	Use pipx install lmapp instead of pip

See Troubleshooting Guide for more.

---

❓ FAQ

Q: How do I install?
pip install lmapp

Q: How do I update?
pip install --upgrade lmapp

Q: Can I use commercially?
Yes! MIT License allows it. See LICENSE.

Q: Does it collect data?
No. 100% local, no telemetry.

More questions? See Troubleshooting Guide.

---

📚 Documentation

• Security Policy
• License
• Changelog

---

📄 License

MIT License - See LICENSE file

This means:

• ✅ Use commercially
• ✅ Modify and distribute
• ✅ Include in closed-source projects
• ✅ Just include the license

Third-Party Licenses

• Ollama: MIT License
• llamafile: Apache 2.0 License
• Pydantic: MIT License
• Pytest: MIT License
• AI Models: Various (see model documentation)

---

🙏 Built With

• Ollama - LLM management platform
• llamafile - Portable LLM runtime
• Pydantic - Data validation
• Pytest - Testing framework
• Meta, Mistral, and other amazing AI model creators

---

⭐ Show Your Support

If lmapp helps you, please:

• ⭐ Star this repository
• 🐛 Report bugs and suggest features
• 📢 Share with friends and colleagues
• 🤝 Contribute improvements
• 📝 Share your use cases

---

📞 Get Started Now

pip install lmapp
lmapp chat

---

📖 Documentation Map

Document	Purpose
QUICKSTART.md	5-minute setup guide ⭐ Start here
docs/installation.md	Installation methods for all platforms
docs/CONFIGURATION.md	Configuration, environment, and settings
docs/development.md	Developer workflow and tips
TROUBLESHOOTING.md	Solutions for common issues
SECURITY.md	Security policy and vulnerability reporting
CHANGELOG.md	Release history
CONTRIBUTING.md	Contribution guidelines
CODE_OF_CONDUCT.md	Community standards
LICENSE	License terms
DEMO.md	Live examples and feature tour
API_REFERENCE.md	Lightweight CLI + HTTP API reference

Additional references:

• docs/ERROR_DATABASE.md - Known errors and fixes

---

Welcome to the future of local AI. 🚀