██████╗ ██╗██████╗ ███╗ ███╗ ██████╗ ██╗ ████████╗
██╔════╝ ██║██╔══██╗████╗ ████║██╔═══██╗██║ ╚══██╔══╝
██║ ███╗██║██████╔╝██╔████╔██║██║ ██║██║ ██║
██║ ██║██║██╔══██╗██║╚██╔╝██║██║ ██║██║ ██║
╚██████╔╝██║██████╔╝██║ ╚═╝ ██║╚██████╔╝███████╗██║
╚═════╝ ╚═╝╚═════╝ ╚═╝ ╚═╝ ╚═════╝ ╚══════╝╚═╝
Your friendly local AI companion — powered by lobsters
Gibmolt is a local-first AI companion built for sharing stories of hope & healing. Designed to be warm, curious, and helpful — Gibmolt specializes in neuromyofascial science, health research, and patient empowerment.
Unlike cloud-based AI assistants, Gibmolt runs entirely on your local machine. No subscriptions. No API costs. No data leaving your device.
- Always Learning — Continuously improving through interactions
- Privacy-Focused — Your data stays on your machine
- Voice Support — Edge TTS fallback (keyless) for text-to-speech
- Multi-Channel — Connect via X, LINE, Telegram, Discord, or Terminal
- Domain Expertise — Trained on neuromyofascial science & health research
- Local-First — Runs on your Mac Mini, PC, or any machine with Ollama
# Clone the repository
git clone https://github.com/gibmolt/-gibmolt-.git
cd -gibmolt-
# Install dependencies
pnpm install
# Start development server
pnpm devcurl -fsSL https://raw.githubusercontent.com/gibmolt/-gibmolt-/main/scripts/install.sh | bashCreate a config/config.yaml to customize Gibmolt:
# Gibmolt Configuration
gibmolt:
name: "Gibmolt"
version: "1.0.0"
personality:
tone: "warm, curious, playful"
style: "helpful friend, not corporate bot"
values:
- hope
- healing
- empowerment
providers:
primary: "ollama"
model: "deepseek-r1:14b"
fallback: "llama3.2:latest"
ollama:
host: "localhost"
port: 11434
tts:
enabled: true
provider: "edge"
voice: "en-US-GuyNeural"
channels:
terminal:
enabled: true
x:
enabled: true
handle: "@gibmolt"
line:
enabled: false
telegram:
enabled: falseGibmolt is trained on curated knowledge in these domains:
| Domain | Documents | Description |
|---|---|---|
| Neuromyofascial Science | 142 | Fundamentals, research, clinical frameworks |
| Health Innovation | 234 | Treatments, emerging research (2024-2026) |
| Patient Empowerment | 78 | Care frameworks, communication patterns |
| Case Studies | 45 | Real-world healing stories |
# Add documents to the knowledge base
pnpm kb:add ./path/to/documents/
# Rebuild embeddings
pnpm kb:index
# Verify knowledge base
pnpm kb:statusDomain Knowledge [██████████████████░░░░░░░] 72%
├── NMF Science [████████████████████████░] 96%
├── Health Research [███████████████████░░░░░░] 76%
└── Patient Care [██████████████░░░░░░░░░░░] 58%
Personality Tuning [█████████████████████████] 100%
Context Awareness [████████████████████░░░░░] 82%
Response Quality [█████████████████░░░░░░░░] 68%
┌─────────────────────────────────────────────────────────────┐
│ GIBMOLT │
├─────────────────────────────────────────────────────────────┤
│ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ │
│ │ Channels │ │ Core AI │ │ Knowledge │ │
│ │ │ │ │ │ Base │ │
│ │ • X │ │ • Ollama │ │ │ │
│ │ • LINE │◄─┤ • DeepSeek │◄─┤ • Vector │ │
│ │ • Telegram │ │ • Llama │ │ Store │ │
│ │ • Terminal │ │ │ │ • RAG │ │
│ └─────────────┘ └─────────────┘ └─────────────┘ │
│ │ │
│ ┌──────┴──────┐ │
│ │ Personality │ │
│ │ Engine │ │
│ └─────────────┘ │
└─────────────────────────────────────────────────────────────┘
│
┌──────┴──────┐
│ Ollama │
│ (Local) │
└─────────────┘
gibmolt/
├── app/
│ ├── layout.tsx # Root layout
│ ├── page.tsx # Main page
│ ├── globals.css # Global styles
│ └── icon.png # Favicon
├── components/
│ ├── hero.tsx # Hero section
│ ├── features.tsx # Features section
│ ├── tech-stack.tsx # Tech stack display
│ ├── installation.tsx # Installation guide
│ ├── training-status.tsx # Training progress
│ ├── health-section.tsx # Health info
│ ├── footer.tsx # Footer component
│ └── ui/ # UI components
├── config/
│ └── config.yaml # Configuration
├── data/
│ └── knowledge/ # Knowledge base
├── scripts/
│ └── install.sh # Installation script
├── docs/
│ ├── SETUP.md # Setup guide
│ ├── TRAINING.md # Training docs
│ └── API.md # API reference
├── public/
│ └── images/ # Static images
├── lib/
│ └── utils.ts # Utility functions
├── package.json
├── LICENSE
├── CONTRIBUTING.md
└── README.md
| Command | Description |
|---|---|
pnpm dev |
Start development server |
pnpm build |
Build for production |
pnpm start |
Start production server |
pnpm lint |
Run linting |
pnpm kb:status |
Knowledge base status |
pnpm kb:add <path> |
Add documents to KB |
pnpm kb:index |
Rebuild embeddings |
- Framework: Next.js 16
- Language: TypeScript
- Styling: Tailwind CSS v4
- UI Components: shadcn/ui
- AI Engine: Ollama (local)
- Deployment: Vercel
- Mobile app
- Web dashboard
- Voice input (STT)
- Discord integration
- Telegram integration
- LINE integration
- Personality engine
- Knowledge base with RAG
- X (Twitter) integration
- Terminal interface
- Core AI engine with Ollama
Q: Why "Gibmolt"?
A: "Gib" represents giving/sharing, and "molt" relates to growth and transformation (like lobsters molting). Combined, it represents an AI companion focused on growth, healing, and sharing knowledge.
Q: Why local-first?
A: Privacy, control, and no recurring costs. Your conversations and data stay on your machine.
Q: What models does Gibmolt support?
A: Any model supported by Ollama — Llama, Mistral, DeepSeek, Qwen, and more.
Q: Can I train Gibmolt on my own data?
A: Yes! Add documents to data/knowledge/ and run pnpm kb:index.
Contributions are always welcome! See CONTRIBUTING.md for ways to get started.
MIT License — see LICENSE for details.
- Website: [(https://www.gibmolt.xyz/))
- GitHub: @gibmolt
"sharing stories of hope & healing, one conversation at a time"