Open sourceWASM-poweredZero data retention

Code to Context.
Instantly.

Generate structured, AI-optimized prompts from any GitHub repository. Feed your entire codebase to ChatGPT, Claude, or Gemini in seconds.

Open Web App Explore CLI

Code never leaves your browserRuns in WebAssembly< 100ms generationPublic & private repos

How it works

Three steps. Zero friction.

Connect your repo

Paste a GitHub URL or sign in to browse your private repositories directly.

Select your files

Pick exactly which files to include. Smart defaults exclude binaries, lock files, and build artifacts.

Generate & copy

The PMP engine runs in your browser via WebAssembly and assembles the prompt locally. Copy or download.

Web Application

Analyze public and private GitHub repositories directly from your browser. Zero installation. The entire prompt is generated locally — your code never touches our servers.

Private repos via GitHub OAuth
WASM engine — runs in your browser
TXT, JSON, XML output formats
Code stays 100% local

Launch Web App

CLI Tool

A Go-based CLI for your local projects. Blazing fast, fully configurable, and scriptable — ideal for CI pipelines and large mono-repos.

Parallel processing (multi-core)
50+ languages & frameworks detected
Security scan — secrets, CVEs, permissions
Git-aware context (--focus-changes)
Token estimation & code quality metrics
Dependency graphs (DOT, JSON, XML)

View on GitHub

Use cases

Built for every AI workflow.

Web + CLI

Code review with AI

Feed a pull-request diff or entire module to Claude or GPT-4. Get architecture-aware feedback in seconds.

CLI

Security audit

Scan for exposed secrets, vulnerable dependencies (CVE database), dangerous permissions and Docker misconfigs.

CLI

Git-focused context

Use --focus-changes to generate a prompt scoped to files modified in recent commits. Perfect for daily stand-ups.

Web + CLI

Dependency mapping

Export DOT/JSON graphs of your module dependencies. Feed them to AI to refactor coupling and spot circular deps.

CLI

Architecture overview

Run --summary-only for AST-based function signatures instead of full source. Condense 10 MB to 50 KB for large mono-repos.

CLI

CI/CD pipelines

Run pmp in GitHub Actions or GitLab CI. Automatically attach an AI-ready snapshot to every PR comment.

Explorer

Browse Any Repo Like an IDE

Navigate files, visualize dependencies, and read code — directly inside the app. No installation required.

benoitpetit / my-projectmy-projectmain

42 files

ready

File treeclick to open

Explorer42

▾src/

▾components/

Button.tsx

Header.tsx

Nav.tsx

▾lib/

utils.ts

api.ts

package.json

tsconfig.json

Dependency graphpan · zoom · collapse

Code viewersyntax · line nums

Header.tsx×

utils.ts

1import { useState } from "react"

3interface Props {

4 title: string

7export function Header({title}: Props) {

8 return (

9 <header className="flex">

10 <h1>{title}</h1>

11 </header>

12 )

13}

TypeScript ReactUTF-8LFLn 9, Col 24

Explorer

Interactive file treecolor-coded by language

Dependency graphpan · zoom · collapse

Syntax-highlighted codeany file · any branch

Open Explorer

Get started in seconds

Install the CLI

Available via Go, cURL, or Homebrew. Transform any local directory into an LLM-ready prompt in milliseconds.

go install github.com/benoitpetit/prompt-my-project@latest

# generate a prompt pmp prompt . --format txt

All installation methods (cURL, Homebrew)

terminal

➜ my-app pmp prompt . --focus-changes --format txt Analyzing project structure... Detected: TypeScript, React, Go, Python (52 files) ⚠ Security: 1 exposed API key (config/dev.env:14) === Analysis Complete === Files processed: 52 (git-changed: 9) Total size: 184 KB Tokens est.: ~46,200 Complexity score: B+ Time: 18ms Output: pmp_output/prompt_20260419.txt

FAQ

Frequently Asked Questions

Is my code safe with Prompt My Project?

Yes. The PMP engine runs entirely in your browser via WebAssembly. Your code is fetched from GitHub, processed locally in your browser, and never stored on our servers. The operator has zero access to your code or generated prompts.

Can I analyze private GitHub repositories?

Yes. Sign in with GitHub OAuth to access your private repositories. Your OAuth token is used only to authenticate API calls and is never shared with third parties.

What output formats are supported?

Prompt My Project supports TXT, JSON, and XML formats for prompts. For dependency graphs, you can export in DOT, JSON, XML, and TXT formats.

How does the WebAssembly engine work?

The core PMP engine is compiled to WebAssembly (WASM) and runs directly in your browser. This means all processing happens client-side with zero server-side computation of your code.

What languages and frameworks does PMP detect?

PMP automatically detects 50+ languages and frameworks including Go, TypeScript, JavaScript, Python, Rust, Java, C/C++, Ruby, PHP, and more. It also identifies dependencies from package.json, go.mod, requirements.txt, Cargo.toml, and other manifests.

How does the security scan work?

The CLI tool includes a built-in security scanner that detects exposed API keys, tokens, and passwords (AWS, GitHub, JWT, SSH keys), checks against a CVE database for known vulnerable dependencies, and audits file permissions and Docker/Nginx configurations for security issues.

What is --summary-only and when should I use it?

The --summary-only flag uses AST-based analysis to generate only function signatures and interfaces instead of full source code. This can reduce a 10 MB codebase to under 50 KB, making it ideal for architecture reviews, documentation, and feeding very large mono-repos to AI assistants with limited context windows.