v2.0.0 — TypeScript rewrite with MCP support

dirgo

Project structure → clipboard → LLM

One command. Zero friction.

Stop manually describing your codebase to AI. One command copies your project structure, dependencies, and configs — ready to paste.

Install Now Star on GitHub See Examples

~/my-saas-app

$ dirgo context

Project: node (npm) [typescript]

├── src/

│ ├── app/

│ ├── components/

│ ├── lib/

│ └── utils/

├── prisma/

├── package.json

└── tsconfig.json

Dependencies (8):

next: ^14.0.0

prisma: ^5.0.0

...

tsconfig.json:

target: ES2022

strict: true

[copied 892B · ~156 tokens]

The Problem

AI can't see your file system

Every time you ask ChatGPT or Claude about your code, you waste time explaining your project structure. Even AI CLI agents still use tree or ls -R — burning tokens on noise. There's a better way.

Method	Time	Friction
Screenshot + paste image	30 sec	High (wastes tokens on pixels)
tree command + manual copy	15 sec	Medium (includes node_modules, .git)
AI agents using tree/ls	5-10 sec	Medium (extra context, no deps)
→ dirgo	< 1 sec	None

Features

Built for the AI-first workflow

⚡

Instant Output

No spinners, no loading bars. Results stream the moment you press enter. Built for speed with Bun.

📋

Auto Clipboard

Output copies automatically. Just Cmd+V into ChatGPT, Claude, or Cursor. Zero extra steps.

🎯

Token Aware

Every output shows estimated tokens. Know exactly how much context you're using before hitting limits.

🔍

Smart Detection

Auto-detects project type, package manager, monorepo structure, and extracts relevant configs.

🌳

Multiple Formats

ASCII tree (default), TOON notation, or JSON. Pick what works best for your LLM or documentation.

🤖

MCP Ready

Built-in Model Context Protocol server. Let AI agents query your project structure programmatically.

Use Cases

Real workflows, real examples

See exactly how dirgo fits into your AI-assisted development.

Use Case #1

Onboarding to a New Codebase

Paste project structure into Claude or ChatGPT and ask "What does this project do?" Get instant architecture overview.

Command: dirgo context

Project: node (npm) [typescript]

├── src/
│   ├── api/
│   ├── components/
│   └── utils/
├── package.json
└── tsconfig.json

Dependencies (12):
  react: ^18.2.0
  next: ^14.0.0
  ...

Use Case #2

Debugging with AI

Show the AI exactly where files are located. No more "I can't see your project structure" responses.

Command: dirgo --focus src/api

└── src/
    └── api/
        ├── routes/
        │   ├── auth.ts
        │   ├── users.ts
        │   └── index.ts
        ├── middleware/
        │   └── auth.ts
        └── client.ts

Use Case #3

Documentation Generation

Generate README sections or architecture docs. The emoji mode makes it documentation-ready.

Command: dirgo -e --depth 2

├── 📁 src/
│   ├── 📁 components/
│   ├── 📁 hooks/
│   └── 📁 utils/
├── 📁 tests/
├── 📦 package.json
├── ⚙️ tsconfig.json
└── 📚 README.md

Commands

Simple CLI, powerful output

Four commands. That's it.

dirgo

Quick tree → copies to clipboard. The default for most tasks.

alias: dirgo tree

dirgo context

Full LLM context: structure + dependencies + configs.

Best for onboarding AI to your project

dirgo deps

Dependencies only. Quick way to show what you're working with.

Parsed from package.json, Cargo.toml, etc.

dirgo serve

Start MCP server for AI agent integrations.

JSON-RPC over stdio

# All the options you might need

--format tree | toon | json

--depth <n> — limit depth

--emoji — add 📁 📄 icons

--focus <path> — expand only subtree

--no-copy — print only

--output <file> — write to file

--include-all — include .git, etc.

--dir <path> — target directory

Language Support

Works with your stack

Parses project manifests to extract dependencies, configs, and workspace structure.

Node.js / TypeScript

package.json, tsconfig.json

Dependencies, scripts, TS compiler options, strict mode detection

Python

pyproject.toml, requirements.txt, Pipfile

Dependencies, dev dependencies, Python version constraints

Go

go.mod, go.work

Module path, Go version, direct and indirect dependencies

Rust

Cargo.toml

Dependencies, features, dev dependencies, workspace members

📦

Monorepo Detection

Automatically detects and lists workspace packages:

npm workspacesyarn workspacespnpm workspacesTurborepoNxLernaGo workspacesCargo workspaces

For AI Developers

Why your AI agent needs dirgo

Building AI-powered dev tools? Here's why dirgo should be in your stack.

Structured, Parseable Output

JSON format enables programmatic processing. No regex needed to extract file paths.

Token Efficient

TOON format uses ~30% fewer tokens than ASCII trees. Every token counts in context windows.

Consistent Interface

Same tool works across Node, Python, Go, Rust. One integration, all project types.

No Hallucination Risk

Real file system data, not guessed structure. Ground truth for accurate code generation.

Model Context Protocol

Built-in MCP Server

Let AI agents query your project structure programmatically. No scraping, no parsing — just clean JSON-RPC.

dirgo_tree

Get directory structure with optional depth, format, and emoji settings

{ "dir": ".", "depth": 2, "format": "tree" }

dirgo_context

Full project context: structure + dependencies + configs

{ "dir": "/path/to/project" }

dirgo_deps

Dependencies only, parsed from project manifest

{ "dir": "." }

# Add to your MCP config (Claude, etc.)

{
  "mcpServers": {
    "dirgo": {
      "command": "dirgo",
      "args": ["serve"]
    }
  }
}

# Example JSON-RPC request

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "dirgo_context",
    "arguments": { "dir": "." }
  }
}

Output Formats

Pick your format

ASCII for humans, TOON for token efficiency, JSON for machines.

Tree (default) ~100 tokens

├── src/
│   ├── api/
│   └── utils/
├── package.json
└── README.md

TOON ~70 tokens

src/
  api/
  utils/
package.json
README.md

JSON parseable

{
  "structure": [
    {"path": "src", ...},
    {"path": "src/api", ...}
  ]
}

Get Started

Install in seconds

# with npm

$ npm install -g @fyzanshaik/dirgo

# with bun (faster)

$ bun add -g @fyzanshaik/dirgo

# then just run

$ dirgo

Works on macOS, Linux, and Windows. Requires Node.js 18+.

View Source npm Package

Your AI assistant
deserves context.

Stop wasting tokens explaining your codebase. Start shipping.

Install dirgo