Overview
What is Flutter Deprecations MCP Server?
Flutter Deprecations MCP Server tracks Flutter API deprecations by scanning Flutter’s GitHub source code for @Deprecated annotations. It provides MCP tools to check code for deprecated APIs and suggest replacements, primarily for AI-assisted Flutter development.
How to use Flutter Deprecations MCP Server?
Install via the provided Makefile (make deps, make build, make run) or build manually with Go. Configure your MCP host (e.g., mcp.json) with the binary path. The server exposes three tools: check_flutter_deprecations, list_flutter_deprecations, and check_flutter_version_info. Command-line flags manage the deprecation cache: --update, --clear-cache, --show-cache, and -vvv for verbose logging.
Key features of Flutter Deprecations MCP Server
- Source-based deprecation tracking from Flutter’s GitHub repository
- Local cache with a 24-hour update interval
- Code analysis for deprecated APIs with replacement suggestions
- Comprehensive scanning of major Flutter directories
- Multi-tier Flutter version detection (CLI, official API, GitHub fallback)
- Multi-platform support including FVM and Docker availability checks
Use cases of Flutter Deprecations MCP Server
- Checking a Flutter code snippet for deprecations during development
- Listing all known Flutter deprecations with replacements and version info
- Querying the latest stable Flutter version and its availability in FVM or Docker
FAQ from Flutter Deprecations MCP Server
What does this server do that alternatives don’t?
It directly scans Flutter’s source code for @Deprecated annotations rather than relying on static changelogs, and it provides MCP tools for real-time queries within AI assistants.
What are the runtime dependencies?
A Go runtime is needed for building; the compiled binary requires no external dependencies. The Flutter CLI is optional but used as the primary method for version detection.
Where are deprecation data stored?
Data is cached locally at ~/.flutter-deprecations/flutter_deprecations.json and automatically refreshed every 24 hours when tools are used.
Are there any known limits?
The server handles GitHub API rate limits gracefully with helpful error messages. The cache prevents excessive API calls; manual updates are available via --update.
What transport and authentication does it use?
The server communicates over standard input/output (stdio) as an MCP server. No authentication is required; it uses public GitHub and Google Storage APIs.
