No server. No account. One binary.
AST Metrics analyzes your codebase (complexity, architecture, coupling, bus factor...) and runs anywhere.
Drop it in any CI. Works offline. Nothing to install, no SaaS, no data leaves your machine.
Fast: 20,000+ lines of code analyzed per second, on a laptop.

Go · PHP · Python · Rust · Java · C# · TypeScript

Documentation | Contributing

AST Metrics analyzing itself. Explore this report live, or try it on any public repository, without installing anything.

Install with Homebrew (macOS, Linux):

brew install ast-metrics/tap/ast-metrics

or with the install script (any platform, downloads an ./ast-metrics binary in the current directory):

curl -fsSL https://install.ast-metrics.dev | sh

Then analyze your project:

ast-metrics analyze --report-html=<directory> /path/to/your/code

Docker image, .deb/.rpm packages and manual downloads: see the detailed installation instructions.

The dependency graph: hubs, natural communities and circular dependencies at a glance.

Read more in the documentation

Run:

# create a .ast-metrics.yaml config file
ast-metrics init 

# Add ruleset to your config file
ast-metrics ruleset add architecture
ast-metrics ruleset add volume
ast-metrics ruleset list # see the list of available rulesets

# Run the linter
ast-metrics lint

You can declare thresholds in your YAML config (Lines of code per method, Coupling, Maintainability...).

Example:

requirements:
  rules:
    architecture:
      coupling:
        forbidden:
          - from: Controller
            to: Repository
          - from: Repository
            to: Service
      max_afferent_coupling: 10
      max_efferent_coupling: 10
      min_maintainability: 70
    volume:
      max_loc: 1000
      max_logical_loc: 600
      max_loc_by_method: 30
      max_logical_loc_by_method: 20
    complexity:
      max_cyclomatic: 10
    golang:
      no_package_name_in_method: true
      max_nesting: 4
      max_file_size: 1000
      max_files_per_package: 50
      slice_prealloc: true
      ignored_error: true
      context_missing: true
      context_ignored: true

This makes it easy to enforce architecture and quality at scale.

Run ast-metrics ruleset list to see the list of available rulesets. Then ast-metrics ruleset add <ruleset-name> to apply a ruleset to your project.

Use the dedicated CI command to run lint and generate all reports in one go:

ast-metrics ci [options] /path/to/your/code

Notes:

• This command runs the linter first, then generates HTML, Markdown, JSON, OpenMetrics and SARIF reports.
• If any lint violations are found, the command exits with a non-zero status but still produces the reports.
• The previous alias analyze --ci is deprecated and will display a warning. Please migrate to ast-metrics ci.

Create a .github/workflows/ast-metrics.yml file in your project with the following content:

name: "AST Metrics"
on:
  pull_request:

permissions:
  contents: read
  pull-requests: write   # optional: allows the action to comment on the pull request

jobs:
  ast-metrics:
    runs-on: ubuntu-latest
    steps:
        - uses: ast-metrics/action-ast-metrics@v2

On each pull request, the action runs ast-metrics review and publishes the result in the check summary and, when permissions allow it, as a single updated comment. On push events it runs a full analysis instead. See action-ast-metrics for all options (fail-on, sarif, html-artifact...).

AI coding agents (Claude Code, Cursor, Copilot...) read code linearly and lack architectural awareness. AST Metrics can act as an MCP server to give them on-demand access to complexity, coupling, dependency graphs, community detection, risk scoring, and test quality — without reading every file.

ast-metrics mcp .

This starts a stdio MCP server exposing 8 tools:

Once configured, just talk to your AI agent naturally. For example:

"What are the riskiest files to refactor?" · "Show me the dependencies of the UserService class — what would break if I change it?" · "Are there complex classes with no tests?" · "I need to work on src/billing/invoice.go, what should I know?"

To use it with Claude Code or any MCP-compatible agent, add a .mcp.json at your project root:

{
  "mcpServers": {
    "ast-metrics": {
      "command": "ast-metrics",
      "args": ["mcp", "."]
    }
  }
}

• ✅ Golang any version
• ✅ Python Python 2, Python 3
• ✅ Rust any version
• ✅ PHP <= PHP 8.5
• ✅ Java any version
• ✅ C# any version
• ✅ TypeScript any version
• 🕛 Flutter
• 🕛 C++
• 🕛 Ruby

AST Metrics is open-source software licensed under the MIT license

AST Metrics is an actively evolving project.

We welcome discussions, bug reports, and pull requests.

➡️ Start contributing here

AST Metrics is built and maintained on free time. If it saved you some of yours:

• ⭐ Star the repository. It costs nothing and it is how most developers discover the tool.
• ❤️ Become a sponsor. Sponsorship directly funds maintenance time and the addition of new languages and metrics.