chrismteryli/DevForge

Local AI workstation. Electron app to run models on your own hardware.

JavaScript

0 commits

updated May 11, 2026

See the code

README

DevForge

Own Your AI. Control Your Costs. Keep Your Data.

The cloud AI providers keep raising prices. APIs get deprecated. Terms change overnight.
DevForge is your insurance policy — a fully local AI workstation that runs on your hardware.

Why Local? • Features • Quick Start • Documentation

Why Local AI?

Cloud AI Problems	DevForge Solution
💸 Rising costs — API prices increase unpredictably	Run models locally for $0/month after hardware
🔒 Data privacy — Your prompts train their models	100% local — nothing leaves your machine
⚡ Rate limits — Throttled during peak usage	Unlimited — your hardware, your limits
📉 Downtime — Service outages halt your work	Always available — works offline
📜 Terms changes — Features removed without notice	You control everything — no surprises
🔌 Vendor lock-in — Stuck in one ecosystem	Model agnostic — swap models freely

DevForge ensures you're never held hostage by a cloud provider's pricing or policy changes.

Features

🏠 Purpose-Built Workspaces

Workspace	Optimized For
Casual	General chat, brainstorming, exploration
Work	Professional tasks, documentation, emails
Code	Development assistance, debugging, code review
Private	Encrypted workspace for sensitive data (journals, notes, personal projects)

🎯 Five View Modes

Interact with AI the way that works best for you:

Stream — Traditional chat flow
Canvas — Visual mind mapping
Document — Collaborative doc building
Timeline — Chronological view
Focus — Distraction-free mode

🤖 Intelligent Model Management

Auto-optimization — Detects model type and applies optimal settings automatically
Hot-swap models — Switch between models instantly
Any GGUF model — Use models from Hugging Face, Ollama, or anywhere
Full parameter control — Temperature, context length, and more

🎨 Local Image Generation

ComfyUI integration — Professional image generation workflow
SDXL & FLUX support — Latest models, running locally
Automatic setup — DevForge can install and configure ComfyUI for you

🔐 Privacy & Security

AES-256 encryption for sensitive workspaces
Zero telemetry — No data collection whatsoever
Offline capable — Full functionality without internet
Quick-hide (Ctrl+Shift+H) — Instant privacy when needed

⚡ Performance

Virtualized lists — Handles thousands of messages smoothly
Lazy loading — Fast startup, load features on demand
Optimized state — Minimal re-renders for smooth UI

Quick Start

Prerequisites

Node.js 18+ — Download
Ollama — Download (free, open source)

Installation

# Clone the repository
git clone https://github.com/jimmyjon121/legendary-bassoon.git
cd legendary-bassoon

# Install dependencies
npm install

# Start Ollama (separate terminal)
ollama serve

# Pull a model (choose based on your hardware)
ollama pull llama3.2        # 3GB - Good for most machines
ollama pull deepseek-coder  # 7GB - Great for coding
ollama pull mixtral         # 26GB - Powerful, needs good GPU

# Run DevForge
npm run dev

Hardware Recommendations

RAM	VRAM	Recommended Models
8GB	4GB	llama3.2, phi3, gemma2:2b
16GB	8GB	mistral, codellama, llama3.1:8b
32GB	12GB+	mixtral, deepseek-coder:33b, llama3.1:70b (quantized)

DevForge's auto-optimization detects your model and adjusts settings for best performance.

Tech Stack

Layer	Technology
Desktop	Electron 31
Frontend	React 18 + Vite
Styling	Tailwind CSS
State	Zustand
Database	SQLite
LLM	Ollama
Images	ComfyUI

Project Structure

devforge/
├── electron/                 # Desktop app backend
│   ├── main.js              # Electron main process
│   ├── preload.js           # Secure IPC bridge
│   └── services/            # Backend services
├── src/                     # React frontend
│   ├── components/          # UI components
│   ├── stores/              # State management
│   ├── services/            # Frontend services
│   └── hooks/               # Custom hooks
└── docs/                    # Documentation

Keyboard Shortcuts

Shortcut	Action
`Ctrl+Shift+H`	Quick hide window
`Ctrl+R`	Reload
`Ctrl+Shift+R`	Force reload
`Ctrl+N`	New conversation
`Ctrl+K`	Model selector
`Ctrl+,`	Settings
`Alt+1-5`	Switch views

Development

npm run dev          # Development mode
npm run build        # Production build
npm run build:win    # Windows installer
npm run build:mac    # macOS installer
npm run build:linux  # Linux AppImage

The Cost Comparison

Usage	Cloud AI (GPT-4)	DevForge
1M tokens/month	~$30-60	$0
10M tokens/month	~$300-600	$0
Heavy daily use	$100+/month	$0
One year	$1,200+	$0

After initial hardware investment, your ongoing cost is just electricity.

Documentation

Contributing

Contributions welcome! Please read the contributing guidelines before submitting PRs.

License

MIT License — Use it, modify it, make it yours.

Stop renting AI. Start owning it.

DevForge — Local AI that works for you, not the other way around.

electron

local-llm

react

Similar repos

hoangnv170752/StudyNest

Local PC app for everyone to use LLM locally 🫡

kartikm7/llocal

Aiming to provide a seamless and privacy driven chatting experience with open-sourced…

171

struklabs/harness-studio

Run open AI models on your own computer — a friendly desktop app and local gateway for llama.cpp…

HilbertraumAI/HilbertRaum

AI, fully local on your computer.

145

AtomicBot-ai/Atomic-Chat

Local AI app and inference engine for agents. Run open-weight LLMs locally — private, 100% offline…

1,812

PurpleDoubleD/locally-uncensored

The all-in-one local AI studio for your desktop: chat, image and video generation and a coding…

2,411

wasifhaq434701-png/Cortex

An Personal Interface for running local LLMs

chorned/frugaLLM-App