OneInfer Edge

OneInfer Edge is an Electron app for managing OneInfer developer workflows from one place: account access, GPU instances, API keys, model routing, local self-hosting, and coding-tool setup.

OneInfer Edge provides one focused workspace for AI infrastructure. It helps developers manage GPU instances, developer keys, credits, models, and routing from a Edge edge app.

What It Does

• Sign in with email OTP
• Manage credits, plans, API keys, routes, and GPU instances
• Deploy local models and expose local inference endpoints
• Build routes across local and cloud targets
• Configure Edge coding tools to use OneInfer models

Product Areas

• GPU Control: launch, inspect, and manage GPU instances.
• Routing Studio: route requests across local and cloud endpoints.
• Quantization: explore model formats, memory, and latency tradeoffs.
• Model Evals: compare quality, cost, latency, and reliability.
• Training & Finetuning: prepare datasets and track model improvement runs.
• Kernel Optimizations: tune inference paths for better hardware efficiency.

For more features check with our wwebsite - https://oneinfer.ai

Authentication

The Edge app uses a developer login flow:

1. Enter the developer email address.
2. Request an OTP from OneInfer.
3. Complete verification to access the workspace.

Tech Stack

Electron, React 18, Vite, TypeScript, CSS, lucide-react, systeminformation, and electron-builder.

Requirements

• Node.js 20+
• npm
• Python only for local model/runtime flows and helper scripts
• Native OS build environment for platform-specific packaging

Setup

npm install
npm run dev

Useful scripts:

npm run lint          # TypeScript checks
npm run build         # Generate assets and build renderer
npm run preview       # Preview Vite build
npm run dist          # Build Edge installer/package
npm run dist:win      # Windows package
npm run dist:linux    # Linux AppImage
npm run dist:mac      # macOS DMG

Config

Default API backend:

https://api.oneinfer.ai/v1

Common environment variables:

• VITE_ONEINFER_API_BASE_URL
• ONEINFER_API_BASE_URL
• CLAUDE_CONFIG_DIR
• OPENCODE_CONFIG_DIR / OPENCODE_CONFIG
• KILO_CONFIG_DIR / KILO_CONFIG

Example .env:

VITE_ONEINFER_API_BASE_URL=https://api.oneinfer.ai/v1
ONEINFER_GOOGLE_Edge_CLIENT_ID=your-google-Edge-client-id
ONEINFER_GOOGLE_Edge_CLIENT_SECRET=your-google-Edge-client-secret

Structure

electron/              Electron main and preload processes
src/                   React app, API client, pages, components, helpers, styles
scripts/               Asset generation and packaging helpers
.github/workflows/     Edge release workflow
developer_*.py         API helper scripts
get_gpus.py            GPU/provider helper script
package.json           scripts, dependencies, Electron Builder config

Packaging

Generated artifacts are written under release/.

• Windows: NSIS installer
• Linux: AppImage
• macOS: DMG

For upload-ready artifact names after a local build:

npm run prepare:upload

Linux packaging should run on Linux or WSL2. When using WSL2, keep the repo under /home/<user>/... for best packaging reliability.