ash310u/awesome-ai-stack

★ 2JavaScriptAudience · vibe coderComplexity · 2/5ActiveLicenseSetup · easy

Mindmap

mindmap
  root((awesome-ai-stack))
    Inputs
      Role choice
      Tool checkboxes
      Client target
    Outputs
      MCP config
      Installed packages
      Merged JSON
    Use Cases
      Set up Cursor
      Wire MCP servers
      Add memory layer
    Tech Stack
      Node
      npx
      TUI

mindmap root((awesome-ai-stack)) Inputs Role choice Tool checkboxes Client target Outputs MCP config Installed packages Merged JSON Use Cases Set up Cursor Wire MCP servers Add memory layer Tech Stack Node npx TUI

Things people build with this

USE CASE 1

Install a curated MCP server set into Claude Desktop without editing JSON by hand

USE CASE 2

Pick a role like Developer or Researcher and let it preselect the right add-ons

USE CASE 3

Bolt on a memory layer such as Mem0 or Zep to an AI assistant

USE CASE 4

Add the UIUX ProMax skill to Cursor in one go

Tech stack

NodenpxTUIJavaScript

Getting it running

Difficulty · easy Time to first run · 5min

Just needs npx and write access to the target client config path on your OS.

MIT license, so you can use, modify, and redistribute it freely as long as you keep the copyright notice.

In plain English

awesome-ai-stack is a small command-line tool that helps you install and wire up extras for AI coding assistants. It runs as a terminal UI, lets you browse a curated catalog of add-ons, and then writes the matching configuration into your client app so the new tools show up automatically. The catalog covers three kinds of things. MCP servers, which are programs that AI assistants can talk to over the Model Context Protocol, include Context7, Browser Use, Firecrawl, GitHub MCP, and Perplexity MCP. Memory layers, which give the assistant a way to remember things across chats, include Claude Mem, Mem0, and Zep Memory. There is also a category called skills, with one entry called UIUX ProMax. The README does not invent extra detail about what each one does beyond naming it. To run it you type npx awesome-ai-stack in a terminal. The flow is short: pick a role such as Developer, Marketer, Researcher, or Designer, or browse all tools; multi-select the packages you want with checkboxes; choose which client you are configuring, which can be Claude Desktop, Cursor, Windsurf, or an install-only mode; then confirm. MCP servers are config-only, meaning the client runs them through npx based on the written config, while memory and skills entries are installed through npx or npm. Config snippets are deep-merged into your existing client config, and the README says existing keys are never overwritten. The tool knows the standard config file locations for each target. For Claude Desktop it writes to the platform-specific path under Application Support on macOS, ~/.config/Claude on Linux, or APPDATA on Windows. For Cursor it writes to ~/.cursor/mcp.json. For Windsurf it writes to ~/.codeium/windsurf/mcp_config.json. For development the README shows the usual clone, npm install, npm run validate to check the registry JSON files, and npm start to launch the TUI. There is an environment variable AWESOME_AI_STACK_USE_LOCAL that forces reading the local packages and roles directories. Contributing a new tool is described as adding a JSON file. The project is MIT licensed and the repository has 2 stars at the time of writing.

Copy-paste prompts

Prompt 1

Run npx awesome-ai-stack and set up Context7 plus GitHub MCP for my Cursor install

Prompt 2

Show me how to add a new MCP server JSON file to the awesome-ai-stack registry

Prompt 3

Use awesome-ai-stack to add Mem0 memory to Claude Desktop on macOS

Prompt 4

Set AWESOME_AI_STACK_USE_LOCAL and test the TUI against my local packages directory

Prompt 5

Help me write a Developer role bundle that installs Firecrawl, Browser Use, and Claude Mem

Open on GitHub → Explain another repo

Generated 2026-05-22 · Model: sonnet-4-6 · Verify against the repo before relying on details.