Personal & Offline AI Assistants That Text
TxtSmith is an innovative MCP-powered chatbot framework designed to run locally on your M4 Mac Mini, providing a completely private AI experience through SMS.
Sustainable technology that respects both privacy and our planet
Meet Smith
Run entirely offline on your M4 Mac Mini with complete privacy. All AI computation is done on your Mac.
- Responds to natural language commands
- Processes all data locally for complete privacy
- Idles around 4 watts, up to 65 watts during inference
Key Features
TxtSmith combines privacy, ease of use, and powerful AI capabilities in one package
Privacy-First Design
AI computation done locally without relying on Big Tech cloud services. Complete control over your data.
SMS Integration
Seamless Twilio integration for natural conversation management through text messages.
MCP Server Ecosystem
Model Context Protocol provides access to specialized tools and capabilities.
Customizable Bot Configuration
Easily configure your bot with characteristics, personality, long-term memory and tools via MCP.
Flexible Architecture
Microservices design with extensibility and configurable language models for maximum flexibility.
Easy Setup
Download, install, and configure with a Twilio phone number in minutes. It's as easy as 'docker compose up' and done!
Technical Specifications
A brief overview of the architecture and system requirements
Architecture Overview
- 1Microservices Design
Modular components for SMS handling, AI processing, tool integration
- 2Docker Containerization
Easy deployment and isolation of service components
- 3Ollama Integration
Efficient management of language models
- 4MCP Protocol
Standardized communication between model and specialized tools
System Requirements
Recommended Hardware
- •Mac Mini with M4 chip
- •16GB+ RAM
- •256GB+ Storage
Software Requirements
- •macOS 14 or newer
- •Docker Desktop
- •Twilio account
System Architecture Diagram
All components run locally on your Mac Mini, ensuring complete data privacy. The modular architecture allows for easy extensibility and customization.
The Team Behind TxtSmith
Meet the experts bringing this innovative solution to life
Jon
Founder & Technical Architect
Jon brings years of AI and system design experience to TxtSmith, specializing in MCP technology and privacy-focused architecture.
Claude
AI & Technical Lead
Claude's expertise in AI integration and model optimization ensures TxtSmith provides intelligent, contextually relevant responses.
Lovable
Frontend Engineer & UX Designer
Focused on creating intuitive interfaces and seamless user experiences, Lovable ensures TxtSmith is accessible and easy to use.
Read more about meFrequently Asked Questions
Common questions about TxtSmith functionality and deployment
Want to collaborate on this project?
We're always looking for passionate collaborators to join us on this journey. If you have ideas, skills, or just enthusiasm to contribute, we'd love to hear from you!
Join our team and help shape the future of AI-powered solutions