Skip to content

About

All in one AI Chat app, with mutiple providers

Topics

Resources

Contributing

Stars

2 stars

Watchers

0 watching

Forks

Repository files navigation

 ██████╗ ██████╗ ███████╗███╗   ██╗███╗   ███╗ ██████╗ ██████╗ ███████╗██╗     ███████╗
██╔═══██╗██╔══██╗██╔════╝████╗  ██║████╗ ████║██╔═══██╗██╔══██╗██╔════╝██║     ██╔════╝
██║   ██║██████╔╝█████╗  ██╔██╗ ██║██╔████╔██║██║   ██║██║  ██║█████╗  ██║     ███████╗
██║   ██║██╔═══╝ ██╔══╝  ██║╚██╗██║██║╚██╔╝██║██║   ██║██║  ██║██╔══╝  ██║     ╚════██║
╚██████╔╝██║     ███████╗██║ ╚████║██║ ╚═╝ ██║╚██████╔╝██████╔╝███████╗███████╗███████║
 ╚═════╝ ╚═╝     ╚══════╝╚═╝  ╚═══╝╚═╝     ╚═╝ ╚═════╝ ╚═════╝ ╚══════╝╚══════╝╚══════╝

The premium, open-source AI playground & chat interface

Multi-provider support. Standalone compilation sandbox. Zero overhead.

🚀 Live Demo: open-models.vercel.app

GitHub Stars License React TailwindCSS PostgreSQL

OpenModels Preview


Why OpenModels?

Most AI chat playgrounds lock you into a single provider or force you to rely on external servers to route your queries. OpenModels gives you absolute control over your key infrastructure. Connect directly to 13+ endpoints, manage sessions client-side without registration, or sign in to persist database history.

  • Zero-Middleman Routing — Connect directly to APIs (NVIDIA NIM, Groq, Gemini) using default keys or bring-your-own credentials.
  • Standalone Development Sandbox — Write HTML/CSS/JS/SVG and compile/preview it in a secure, isolated container instantly.
  • Secure Encrypted Storage — User-provided API keys are encrypted at rest using industry-grade AES-256-GCM on the backend.

Two components, one repository

Component Path Technical Stack
React Client client/ React 19, Vite 8, Zustand (Sliced store), Monaco Editor, TailwindCSS v4
Express Server server/ Node.js, Express 4, Prisma (PostgreSQL), AES-256-GCM, Server-Sent Events (SSE)

Features

  • 🔌 13+ Built-in Providers — Seamlessly query NVIDIA NIM, Gemini, Groq, OpenRouter, Mistral, GitHub Models, Together, SambaNova, Cerebras, Cohere, Cloudflare Workers, Z.AI, and Agens AI.
  • 🛠️ Standalone Compilation Sandbox — Run HTML, CSS, JavaScript, and SVG codes directly within an inline frame. Includes Monaco Editor, hot reloading (800ms debounce), and client-side downloads.
  • 🔍 Web Search & Deep Think — Enhance queries with real-time web search results (powered by Firecrawl) and Deep Thinking/Reasoning models (e.g., DeepSeek R1).
  • 🔒 AES-256-GCM Encryption — Secure API keys stored in user settings. Keys are encrypted with a unique IV and authentication tag before insertion.
  • 👤 Guest Session Persistence — Chat instantly without signing up. Guest sessions save up to 5 messages to local state/IndexedDB.
  • 🗂️ Message Versioning & Editing — Edit past user prompts, navigate through alternative generation branches, and trigger hot regenerations.
  • 🎭 System Personas — Seed custom personas or utilize default system presets (seeded on startup) to guide AI responses.
  • 📱 Responsive Layout — Compact headers, scroll-independent panels, and responsive navigation drawers designed for viewports down to 320px.

Quick start

# Clone the repository
git clone https://github.com/krishcodes07/openmodels.git
cd openmodels

# Install root dependencies
npm install

# Setup environment variables
cp .env.example .env
# Edit .env with your DATABASE_URL, JWT_SECRET, and ENCRYPTION_KEY

# Initialize database schema (Prisma)
cd server
npm install
npx prisma generate
npx prisma db push

# Start both client and server concurrently
cd ..
npm run dev

Supported Providers & Capabilities

Provider ID Vision Deep Think Web Search Recommended Env Key
NVIDIA NIM nvidia ✅ ✅ ✅ NVIDIA_API_KEY
Groq groq ❌ ✅ ✅ GROQ_API_KEY
Google Gemini gemini ✅ ✅ ✅ GEMINI_API_KEY
OpenRouter openrouter ✅ ✅ ✅ OPENROUTER_API_KEY
Mistral AI mistral ✅ ❌ ✅ MISTRAL_API_KEY
GitHub Models github ✅ ✅ ✅ GITHUB_API_KEY
SambaNova sambanova ❌ ❌ ✅ SAMBANOVA_API_KEY
Cerebras cerebras ❌ ❌ ✅ CEREBRAS_API_KEY
Cloudflare AI cloudflare ❌ ❌ ✅ CLOUDFLARE_API_TOKEN
Cohere cohere ❌ ❌ ✅ COHERE_API_KEY
Z.AI zai ✅ ❌ ✅ ZAI_API_KEY
Agens AI agnes ❌ ❌ ✅ AGNES_API_KEY

Permissions & Sandboxing

Secure execution by default. OpenModels runs all compiled web previews inside an isolated <iframe> with strict security constraints:

<iframe srcDoc={srcDoc} sandbox="allow-scripts allow-modals" />

By omitting the allow-same-origin flag, the compiled preview cannot access your host cookies, localStorage, IndexedDB, or session tokens, completely isolating potential XSS vulnerabilities from the main application.


Documentation

Topic Link
System Architecture docs/architecture.md
Local Setup Guide docs/getting-started.md
Adding a Custom Provider docs/provider-guide.md
Interactive Sandbox Engine docs/sandbox-preview.md
Authentication & Credentials Encryption docs/auth-and-security.md
Production Deployment Manual docs/deployment.md
Contributing CONTRIBUTING.md

License

This project is licensed under the MIT License.

About

All in one AI Chat app, with mutiple providers

Topics

Resources

Contributing

Stars

2 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages