Back-end developer, AI automation engineer, and SaaS founder with 15+ years of shipping products that work.
My first SaaS was acquired for over 10x ARR multiple. Currently running ZenMic — an AI Podcast Generator that has powered 5,000+ episodes for creators worldwide.

• 📍 Jaipur-based indie hacker | #BuildInPublic
• 🔧 15+ years shipping full-stack web products from idea to production
• 🤖 Deep focus in 2025–2026 on autonomous AI agents and serverless automation pipelines
• 🚀 Bootstrapped → acquired → building again
• 💬 Follow the journey: @motyar on X

"I don't just write code. I solve problems that happen to require code."

From idea to published podcast episode in seconds. No mic, no studio, no editing.

ZenMic converts any text, topic, or blog post into a ready-to-publish audio podcast using AI — multi-voice conversations, Neural TTS, instant publishing, and a built-in SEO flywheel.

• ✅ Topic-to-podcast in under 3 minutes
• ✅ Multi-voice AI conversations (male + female Neural voices)
• ✅ Public REST API with key management
• ✅ Built-in SEO & marketing intelligence pipeline
• ✅ Fully automated content marketing — runs without manual intervention

🔗 zenmic.com · Private repo — happy to demo the architecture on a call

💰 SaaS Exit: 10x+ ARR Multiple
Built and sold my first bootstrapped SaaS — acquired for over 10x ARR multiple. Proof that simple tools solving real niche pains can produce outsized outcomes.

🤖 Autonomous AI Systems Pioneer
In 2025–2026 I've been systematically building and publishing open-source AI agent frameworks — from self-improving codebases to personal AI assistants to automated business research pipelines — all running on zero-infrastructure GitHub Actions.

🔄 Serial Builder
From jQuery plugins used on thousands of websites, to Go REST APIs forked by hundreds of developers, to a live AI SaaS — a consistent track record of shipping things people actually use.

📊 Open Source Impact
320+ GitHub stars · 47+ repositories · Code running on 5 continents.

This is where my head has been in 2025–2026. Every project below runs on zero servers — pure GitHub Actions.

Your codebase improves itself. 24/7. No human required.

An autonomous agent loop powered by Google Gemini and GitHub Actions. Every scheduled run picks the next pending task from a work queue, executes it using specialized AI skills (engineering, research, marketing, analysis), saves results, and automatically queues the next task — creating a self-sustaining development loop. Fork it, point it at your project mission, and let it ship.

• ✅ Fully self-directing: reads mission → executes → plans next step → repeats
• ✅ Role-aware agents: CEO (strategy), Engineering, Marketing, Research
• ✅ Collective memory that grows with every run
• ✅ Zero infrastructure — just GitHub Actions + Gemini API

A self-aware AI that lives in your GitHub repo, talks on Telegram, and gets smarter over time.

GitButler is a personal AI assistant powered by GPT-4o-mini that runs entirely on GitHub Actions, communicates through Telegram, and maintains persistent memory via Git commits. It knows who you are, remembers your preferences, and can even create GitHub issues/PRs to improve its own code. Supports continuous mode (~10 second response time) with auto-sleep to conserve minutes.

• ✅ Persistent identity system: soul, memory, user profile, all in markdown files
• ✅ Continuous mode: stays active up to 6 hours with ~10 second responses
• ✅ Self-improving: creates issues and PRs to modify its own codebase
• ✅ Extensible skills system — add new capabilities with a single markdown file

A Telegram bot that runs on GitHub Actions and can execute real shell commands from natural language — "show disk usage", "list files", "create a directory" — while also maintaining persistent memory across sessions. Built entirely in shell script with jq.

• ✅ Natural language → shell command execution (with safety guardrails)
• ✅ Three-layer memory: reflections (soul.md), events (memory.md), user profile (user.md)
• ✅ Self-updating memory committed back to the repo automatically

Daily competitive intelligence — free, automated, committed to your repo.

Scrapes HackerNews and Reddit for pain points, performs Google Autocomplete keyword expansion (alphabet-soup method), scores keywords by opportunity, and commits timestamped JSON/CSV/Markdown results to the repo every day at 6 AM UTC. No dashboard, no VPS, no ongoing cost.

• ✅ HackerNews trending topics + Reddit community pain-point scraper
• ✅ Long-tail keyword research via Google Autocomplete (no API key needed)
• ✅ Opportunity scoring (0–4) and SQLite keyword database
• ✅ Configurable niche, subreddits, and seed keywords — outputs to output/

A serverless Telegram bot that runs entirely inside GitHub Actions — no server, no Docker, no npm install. Uses GPT-4o-mini for natural language understanding, manages GitHub issues & PRs, and remembers conversation history. Extensible via a clean skill-registry pattern — add a new capability by adding one object to an array.

Fully automated LinkedIn PDF carousel generator — GPT-4o-mini generates ideas and content daily via GitHub Actions, renders professional PDFs with 5 colour themes, delivers to Telegram for one-tap posting. Zero manual effort after setup.

Transcribes audio and generates TikTok/Reels-style MP4 with animated subtitles — runs entirely on GitHub Actions with zero API costs. Whisper (local) + FFmpeg + three subtitle modes: chunk, word, karaoke.

A full-featured, enterprise-grade Loyalty Management System backend built with NestJS (TypeScript) and PostgreSQL — the kind of thing you'd hire a team to build.

• ✅ Multi-tenant architecture with strict data isolation per client
• ✅ JWT authentication + API key auth for third-party integrations
• ✅ Role-based access control: SuperAdmin, ClientAdmin, ClientStaff, Customer
• ✅ Campaigns, loyalty rules, points calculation, redemption with race-condition prevention
• ✅ Swagger/OpenAPI docs, rate limiting, full audit trail
• ✅ Stack: NestJS · TypeScript · PostgreSQL · TypeORM · Passport · Docker

Demonstrates the kind of clean, scalable backend architecture I bring to client engagements.

Plug-and-play jQuery plugin for smooth firefly particle effects. Used by WordPress themes and websites worldwide — including a Yootheme template.

158 ⭐ · 56 forks

Clean, production-ready full CRUD REST API in Go backed by MySQL. A go-to starter template for developers — one of my most-forked repos.

55 ⭐ · 19 forks

Lightweight real-time chat using HTML5 Server-Sent Events and TailwindCSS. Push-based server communication with no WebSocket overhead.

26 ⭐ · 8 forks

Turn any Markdown file into a Reveal.js presentation with themes, transitions, and a live editor. No install — paste and present. Live at motyar.github.io/markshow.

3 ⭐ · Topics: markdown markdown-editor revealjs

• 🧱 "We need an MVP built fast" → I architect and ship from zero to production. No hand-holding required.
• 🤖 "We want AI features" → I've shipped GPT, Gemini, Whisper, Azure TTS integrations — and I build autonomous pipelines, not just API wrappers.
• ⚙️ "We need an AI agent or automation pipeline" → This is my current specialty. Autonomous loops, self-improving systems, serverless bots.
• 🏢 "We need an enterprise-grade backend API" → Multi-tenant, role-based, documented, tested. See lms-test as an example.
• 🔌 "We need a solid REST API" → Built them in Go, NestJS, Node.js, and PHP at multiple scales.
• 📈 "We need content/marketing automation" → Fully automated pipelines running 24/7 with zero ops overhead are my thing.
• 🚀 "We need someone to own a project end-to-end" → That's exactly what I do. I deliver working software, not tickets.
• ⚙️ "Our codebase is a mess" → Technical audit, refactor, cleanup. I've seen it all.

I take on a limited number of engagements at a time to ensure every client gets real attention and real results.

→ Reach out on X: @motyar
→ Or open an issue on this repo

I respond to every serious inquiry within 24 hours.