# kannajune/mcp-architect [Health: Active]

**Category:** 🔒 Delivery  
**Repository:** https://github.com/kannajune/mcp-architect  
**GitHub Stars:** 3  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/kannajune-mcp-architect

## Description
Gives any AI assistant real architectural understanding of a codebase: tech-stack overview, internal dependency graph with cycle detection, risk hotspots, and module summaries. Local, zero-config, no API keys. uvx mcp-architect.

## Tools
Capabilities this server exposes over MCP:

- **architecture_overview** — Languages, frameworks, ecosystems, size, top‑level structure, entry points
- **dependency_graph** — Internal import graph, architectural hubs, **circular dependencies
- **impact_analysis** — What breaks if you change X** — direct importers + transitive blast radius, hub risk
- **hotspots** — Largest / most complex / most‑changed (git) / highest‑risk files
- **explain** — Deep‑dive a folder or file: classes, functions, external deps

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `uvx` (confidence: high):

```json
"mcpServers": {
  "mcp-architect": {
    "command": "uvx",
    "args": ["mcp-architect"]
  }
}
```

## Documentation

## What the kannajune/mcp-architect MCP server does

The kannajune/mcp-architect MCP server gives an MCP client structured information about a local repository instead of requiring an assistant to infer architecture from a pasted file tree. It can identify languages, frameworks, ecosystems, repository size, top-level directories, and likely entry points. It also describes internal module relationships and provides summaries of the classes, functions, and external dependencies found in a selected file or folder.

The analysis is intended for architectural orientation and coding decisions. For example, an assistant can inspect which modules depend on a target before proposing a change, or find files that deserve review because they are large, complex, frequently changed, or otherwise risky.

## How it works

The server runs on the developer's machine and reads the selected codebase locally. It does not send source code over the network, use telemetry, call an external API, or require an LLM. Python files receive full AST-based analysis. JavaScript and TypeScript imports are parsed for dependency information, while file and line-count statistics are available across more than 25 languages.

Its dependency and complexity results are heuristic. They are designed to provide fast situational awareness rather than replace a dedicated static-analysis system. The analysis package can also be imported independently as a Python library, including functions such as `get_overview` and `get_dependency_graph`.

## Setup and configuration

Install the package with `pip install mcp-architect`, or let an MCP client retrieve and run it with `uvx mcp-architect`. A client configuration needs a server entry using the `uvx` command and `mcp-architect` as its argument. The README specifically documents compatibility with Claude Desktop, Cursor, Windsurf, and Cline.

Set `MCP_ARCHITECT_ROOT` in the server environment when tools should default to one fixed repository and callers should be able to omit paths. Otherwise, tools can receive repository paths directly. No API key, model download, or additional credential is required.

## Tools and capabilities

The kannajune/mcp-architect MCP server exposes these MCP tools:

- `architecture_overview` reports languages, frameworks, ecosystems, repository size, top-level structure, and entry points.
- `dependency_graph` builds an internal import graph, identifies highly depended-upon modules, and checks for circular dependencies.
- `impact_analysis` lists direct importers, transitive change reach, and hub-related risk for a selected target.
- `hotspots` ranks files by size, complexity, Git change history, and calculated risk.
- `explain` examines a file or folder and summarizes its classes, functions, and external dependencies.

Together, these tools support repository orientation, dependency review, change-impact questions, and prioritization of files for inspection.

## Limitations and notes

The documented import-graph support is strongest for Python, JavaScript, and TypeScript. The roadmap lists Go, Rust, and Java import graphs as future work, so those languages should not be assumed to have equivalent dependency analysis. Layered-boundary checks, architecture comparisons, and optional local-LLM narratives are also listed as planned rather than current capabilities.

The project is released under the MIT license. Its runtime is described as using the MCP SDK plus Python standard-library analysis components, and the repository targets Python 3.10 or newer.

_Full upstream README: https://allmcps.com/mcp/kannajune-mcp-architect/readme_