I'm 11 years old and I built an AI agent platform with 250+ tools

Mohab Ahmed — Sun, 22 Mar 2026 07:37:16 +0000

Hey Dev.to community 👋

My name is Mohab. I'm 11 years old and I'm from Egypt 🇪🇬

I want to share something I built and get honest feedback from real developers.

The Problem

Every time I tried to use AI agents, the setup was painful. Different providers have different APIs, different formats, different quirks. There was no single place that unified everything. You had to jump between 5 different docs just to get started.

I thought: someone should fix this. Then I thought: why not me?

What I Built

I built VoiceDev — an open source AI agent platform that unifies everything into one interface.

Here's what it includes:

🔧 250+ Tools

Real working implementations across every category a developer needs:

FileSystem — read, write, copy, move, hash, watch, compress
Shell — exec, pipe, background processes, cron jobs, Docker
Git — clone, commit, push, merge, rebase, stash
Browser — navigate, screenshot, extract, click, automate
Remote — SSH, RDP, VNC, WinRM, Wake-on-LAN
Security — port scan, SSL check, XSS test, SQL injection test
Voice — TTS and ASR across multiple providers
Channels — Telegram, WhatsApp, Discord, Slack, Email, MCP

🎨 123+ Skills

Skills are multi-step workflows that combine multiple tools. For example a "deploy" skill might run tests, build the project, commit, push, and notify your Slack — all in one command.

🤖 15 AI Providers

OpenAI, Anthropic, Google, DeepSeek, Groq, Mistral, xAI, ElevenLabs, Moonshot, MiniMax, GLM, Cohere, Together, Qwen, Replicate

📡 80+ Verified Models

All models verified against official provider docs as of March 2026:

GPT-5.4
Claude Sonnet 4.6
Gemini 3.1 Pro
Grok 4.20 Beta
Llama 4 Maverick
DeepSeek V3.2-Exp
And many more

🎤 Voice Pipeline

TTS: GPT-4o Mini TTS, ElevenLabs V3 (70+ languages), MiniMax Speech 2.6
ASR: Whisper Large V3 Turbo (216x realtime via Groq), ElevenLabs Scribe V2

🖥️ Desktop App

Built with Tauri v2 + Rust. Compiles to native installers for Windows (.msi), macOS (.dmg), and Linux (.deb, .AppImage).

🔒 5-Layer Security

Static code analysis
Sandboxed execution
Granular permission system
Rate limiting
Audit logging

Tech Stack

Frontend: Next.js 15 + TypeScript
Desktop: Tauri v2 + Rust
Database: Prisma
Browser automation: Playwright
License: MIT — free forever, no telemetry, no paywalls

Quick Start

git clone https://github.com/Mohabsmar/voicedev.git
cd voicedev
npm install
npx prisma generate
npx prisma db push
npm run dev

What I Learned

Building this taught me more than any tutorial ever could:

How AI provider APIs actually work under the hood
Why security layers matter (I added 5 of them after reading about real exploits)
How to structure a large TypeScript codebase
That shipping something imperfect is better than never shipping

Why I'm Sharing This

I'm not sharing this to show off. I'm sharing because:

I want real feedback from experienced developers
I want other young developers — especially in Egypt and across Africa — to know that age is not a barrier
I genuinely believe open source makes the world better

If you find a bug, open an issue. If you have feedback, leave a comment. If you think something is wrong with my approach, tell me directly. I can handle it.

Links

⭐ GitHub: https://github.com/Mohabsmar/voicedev

Thank you for reading.

— Mohab, 11, Egypt 🇪🇬

"Technology doesn't care about your age. It only cares about what you build."

DEV Community: Mohab Ahmed