flutter_architect_mcp is a Model Context Protocol (MCP) server and command-line utility designed to automate the process of bootstrapping new Flutter applications. It sets up architectural structures, injects required package dependencies, sets up routing, databases, networks, and backend configurations, and copies a comprehensive collection of shared utilities, widgets, and extensions into the generated codebase.
Audit architecture, security, memory, and code quality in Flutter, Laravel, Node, Python, & Vue.
2026-11-09
ดาว
—
เครื่องมือ
—
อัปไทม์
—
What is Provided
The project contains a complete template library and a generator pipeline organized as follows:
--target: Absolute path of parent folder (defaults to current path)
3. Direct Programmatic Script
Run the pre-configured flutter_architect_mcp.dart script to quickly test or launch a default workspace setup:
dart run bin/flutter_architect_mcp.dart
This script generates a pre-configured project named expense_tracker_with_mcp with clean architecture, bloc, hive, firebase, dio, and go_router in the local directory.
Runtime Performance Check Tool
The Runtime Performance Monitor spins up a local server hosting a live web-based DevTools-like dashboard to profile active screens, HTTP API timing latencies, heap memory (RAM), and local database storage consumption. Stopping a session compiles all metrics into a formatted printable PDF HTML report.
1. Launching the Dashboard Server
You can launch the dashboard server directly from CLI:
dart run lib/tools/scanners_and_audit_reports/run_runtime_perf.dart --port 8080
Port Flag: Set custom ports using --port <number> (defaults to 8080).
Access Dashboard: Open http://localhost:8080 in your web browser.
2. Registering in MCP Server
You can start the tool via the MCP server using the start_runtime_perf_check tool:
Copy the helper code integrations displayed under the Integration Guide tab of your dashboard:
Network Request Interceptor: Add the custom Dio interceptor to your HTTP client to log request methods, endpoints, response statuses, and durations.
Screen Route Observer: Register the PerfNavigatorObserver in your MaterialApp observers to track navigation changes.
Periodic Memory & Storage Streamer: Launch the background timer inside main() to stream RAM (RSS) size and local DB file size points.
All data streams directly to the dashboard, and clicking the Stop & Save PDF button on the dashboard will export a detailed engineering review report with remediation steps!
4. Connecting Simulators, Emulators, & Real Devices
To allow your running mobile app to post metrics back to the local performance server, configure the connection URL according to your target device environment:
A. iOS Simulator (macOS host)
Since the iOS simulator runs directly on the host Mac, it can connect via localhost:
Connection URL:http://localhost:8080/api/record
B. Android Emulator (Virtual Device)
Android emulators run within a virtual loopback environment. To access the host machine's localhost, use the special IP address 10.0.2.2:
Connection URL:http://10.0.2.2:8080/api/record
C. Physical Android Device (Connected via USB)
You can forward port 8080 from the device back to your computer using ADB:
Run the following command in your terminal:
adb reverse tcp:8080 tcp:8080
Your app can now use localhost directly:
Connection URL:http://localhost:8080/api/record
D. Physical iOS / Android Device (Connected via Wi-Fi)
Ensure both your development computer and mobile device are connected to the same Wi-Fi network.
Find your computer's local IP address (e.g. 192.168.1.45 or 10.0.0.12).