kevinkamto /
deep-thinker
A full-stack AI research agent that breaks complex questions into parallel research threads, searches the web, summarizes findings, and synthesizes a final report - all streamed live to the UI.
Loading repository data…
nitrocloudofficial / repository
The full-stack TypeScript framework to build, test, and deploy production-ready MCP servers and AI-native apps.
Documentation · Quick Start · Blog · NitroStudio · Discord
npx @nitrostack/cli init my-server
cd my-server
npm install
npm run dev
Your MCP server is running. Connect it to any MCP-compatible client.
Once your project is scaffolded, open the same folder in NitroStudio for visual testing and debugging.
my-server project folderBuilding MCP servers today means stitching together boilerplate, reinventing authentication, and hoping your tooling scales. NitroStack gives you an opinionated, batteries-included framework so you can focus on what your server actually does.
import { McpApp, Module, ToolDecorator as Tool, z, ExecutionContext } from '@nitrostack/core';
@McpApp({
module: AppModule,
server: { name: 'my-server', version: '1.0.0' }
})
@Module({ imports: [] })
export class AppModule {}
export class SearchTools {
@Tool({
name: 'search_products',
description: 'Search the product catalog',
inputSchema: z.object({
query: z.string().describe('Search query'),
maxResults: z.number().default(10)
})
})
@UseGuards(ApiKeyGuard)
@Cache({ ttl: 300 })
@Widget('product-grid')
async search(input: { query: string; maxResults: number }, ctx: ExecutionContext) {
ctx.logger.info('Searching products', { query: input.query });
return this.productService.search(input.query, input.maxResults);
}
}
One decorator stack gives you: API definition + validation + auth + caching + UI — zero boilerplate.
NitroStack is modular. Install only what you need:
The implementation workspace for NitroStack packages lives in typescript/.
| Package | What it does | Install |
|---|---|---|
@nitrostack/core | The framework — decorators, DI, server runtime | npm i @nitrostack/core |
@nitrostack/cli | Scaffolding, dev server, code generators | npm i -g @nitrostack/cli |
@nitrostack/widgets | React SDK for interactive tool output UIs | npm i @nitrostack/widgets |
NitroStudio is a desktop app purpose-built for developing MCP servers. Open your project folder — it handles the dev server for you.
Real-time tool testing Execute tools, inspect payloads, and debug request/response cycles.
Built-in AI chat Talk to your MCP server through an integrated AI assistant.
| Resource | Description |
|---|---|
| Getting Started | Installation, quick start, and first project |
| Server Concepts | Modules, DI, and architecture deep dive |
| Tools Guide | Defining tools, validation, annotations |
| Widgets Guide | Building interactive UI components |
| Authentication | JWT, OAuth 2.1, API key setup |
| CLI Reference | All CLI commands and options |
| Deployment | Production checklist, Docker, cloud platforms |
We welcome contributions of all kinds — bug fixes, features, docs, and ideas. Read the Contributing Guide to get started.
Looking for a place to begin? Check out issues labeled good first issue.
NitroStack is open-source software licensed under the Apache License 2.0.
Selected from shared topics, language and repository description—not editorial ratings.
kevinkamto /
A full-stack AI research agent that breaks complex questions into parallel research threads, searches the web, summarizes findings, and synthesizes a final report - all streamed live to the UI.
anwitac246 /
Campaign-AI is a full-stack AI-powered marketing assistant designed to empower small businesses and marketing teams. It streamlines the entire campaign creation process by leveraging a crew of specialized AI agents for market research, strategy development, copywriting, and image generation.
Poojan38380 /
AI-powered browser IDE built with Next.js 16, Gemini 2.5 Flash, Inngest Agent Kit, Convex, and WebContainers. Describe what you want → AI reads, writes & runs your code live in the browser.
Nexa is an intelligent agentic AI system that transforms natural language prompts into fully functional full-stack applications. It enables users to describe what they want to build, and the system automatically generates structured backend logic, frontend UI, and integrated workflows using modern AI orchestration techniques.
KeepSerene /
Writ — a full-stack, terminal-based agentic AI coding assistant. A pnpm monorepo containing the Bun/opentui TUI client, a Hono API server, a Prisma/Neon database layer, and shared schemas.
AralAutomata /
Full Stack Agentic AI Pipeline for Educational Use: Deno runs the LangChain/LLM agent backend + Bun runs the Next.js UI + SQLite database for audit and history + Per-entity memory and immutable snapshots logged as JSON provide continuity across run cycles + TypeScript Fun Time