Creator Toolkit
A personal project simulating an AI toolkit for content creators—includes voice cloning, thumbnail editing, research agents, and subtitle generation. Built to learn full-stack + AI integration hands-on.
What This Is
A micro SaaS toolkit for YouTubers and content creators, built as a hands-on learning project. It packs the entire content pipeline into one workspace: thumbnail editing, voice cloning, research, reel generation, and subtitles. Built solo over 14 weeks to master full-stack development and AI integration.
Features
Thumbnail Generator + Image Editor A custom image editor built with React and Konva.js. Layered composition, drag-resize, real-time preview. AI inpainting for adding or removing elements. Go from template to hand-built to auto-generated thumbnails.
Voice Toolkit
- Voice cloning with ElevenLabs
- Multilingual dubbing and style changes on the fly
- Switch from narrator to character voices instantly
Research Wizard Powered by OpenAI and Perplexity. Auto-summarizes topics, generates outlines, suggests titles and tags. Exports research packs as PDFs.
Reel Generator Extracts the best 15-60 second clips from long videos for Shorts and Reels. Suggests hooks and previews. One-click repurposing.
Subtitle Generator Whisper-driven audio-to-text with timestamping. Planned: styled subs and baked-in captions.
Technical Highlights
Queue-Based Media Processing Heavy video and audio work gets offloaded to Bull queues with Redis. FFMPEG handles conversion, slicing, and transformation without blocking user requests. The main app stays responsive while jobs run in the background.
Real-Time Updates with SSE Server-Sent Events push instant status updates to the frontend. Users see when their video or audio task finishes without refreshing.
Modular AI Orchestration 5+ AI APIs working together: OpenAI (GPT/Whisper), Stable Diffusion, ElevenLabs. Each service is swappable. Rate limits and failures are handled gracefully.
Tech Stack
Frontend: Next.js, React, Tailwind CSS, react-konva, Redux Toolkit, React Hook Form, Radix UI, Stripe/PayPal
Backend: NestJS, MongoDB, Redis, Bull queues
AI/Media: OpenAI, Whisper, Stable Diffusion, ElevenLabs, FFMPEG, Puppeteer
Infra: Docker, AWS EC2, GitHub Actions
What I Learned
- Designing AI as the backbone, not a bolt-on feature
- Managing complex multi-stage media pipelines without breaking the UX
- Building custom canvas-driven editors with real-time interaction
- Orchestrating multiple APIs, queues, and rate limits in production
- Keeping every component modular and swappable for future growth
