Appnest App MCP

Model Context Protocol (MCP) server for Appnest App projects: scaffold base code, set up AI context, validate runtime prerequisites, and run the appnest-development-engine CLI workflows.

---

Requirements

Requirement	Details
Node.js	22+ (matches appnest-development-engine / Appnest app; uses fetch, ESM).
MCP client	Any client that launches MCP over stdio (e.g. Cursor MCP config).
appnest-development-engine	On PATH, for run_appnest_command and for hi appnest (precheck, install-packages). Use setup_appnest_app_runtime to verify Node 22+ and install appnest-development-engine globally if missing.
Network	Clone/setup tools download zips from GitHub (appnest-sample-basecode, appnest-sample-tools).
Project folder	Pass explicit paths where tools support them (e.g. targetDir, workingDirectory).

Environment variables

Variable	Purpose
.env	Loaded via dotenv (e.g. future API keys).

Cursor / MCP config example

In your Appnest project root, create .cursor/mcp.json with the following:

{
    "mcpServers": {
        "appnest-app-mcp": {
            "command": "npx",
            "args": [
                "-y",
                "@sparrowengg/appnest-app-mcp"
            ],
            "transport": "stdio"
        }
    }
}

Same file via shell:

mkdir -p .cursor && cat << 'EOF' > .cursor/mcp.json
{
    "mcpServers": {
        "appnest-app-mcp": {
            "command": "npx",
            "args": [
                "-y",
                "@sparrowengg/appnest-app-mcp"
            ],
            "transport": "stdio"
        }
    }
}
EOF

---

Tools

Tool	Purpose
hi_appnest	Full onboarding: setup basecode → setup_appnest_ai_context → precheck → install-packages.
setup_appnest_app_basecode	Download & extract official base zip into targetDir (optional; when omitted, uses the server host process working directory). Skips if the folder isn't empty (allowed: .cursor, .DS_Store, Thumbs.db) unless force: true.
setup_appnest_ai_context	Replace appnest-ai-context/ (delete if present), then download & extract from the sample tools repo (appnest-sample-tools zip).
create_appnest_prd_prompt	Reads appnest-ai-context/appnest-prd-generator/prd-generation-workflow.md, tags its path, embeds the file, and appends samplePrd for the AI editor (requires AI context setup first).
setup_appnest_app_runtime	Ensures Node.js 22+ (on macOS/Linux can install via nvm + official install script if MCP is on older Node), global appnest-development-engine, then appnest-development-engine app help. Needs network when installing.
run_appnest_command	Runs appnest-development-engine app + subcommand with optional workingDirectory. Subcommands: init, precheck, install-packages, start, pack, validate, ai-context. Discover: appnest-development-engine app, appnest-development-engine app help, or appnest-development-engine app -h.

---

Sample prompts

Use natural language in your MCP-enabled client; the assistant can map these to hi_appnest, setup_appnest_ai_context, create_appnest_prd_prompt, and related tools. Replace placeholders with your real paths and product idea.

Full onboarding (basecode + AI context + precheck + install-packages):

Hi Appnest — do the base setup.

Refresh AI context only (after onboarding or when templates change):

Update the Appnest AI context.

Generate a PRD-style prompt (requires AI context first):

Generate an Appnest app PRD for this idea: an Appnest app that <describe your app — e.g. integrates with our product and automates X>.