Open Source Technologies

Open-source text-to-image diffusion model by Black Forest Labs. Generates every AI image, sticker, icon, and banner across the site.

Google DeepMind's open-source large language model. Powers text generation, paraphrasing, summarization, OCR, image captioning, and every writing tool — in the browser on capable devices and on-server as fallback.

Trimap-free portrait matting model for real-time background removal without pre-processing. Runs in the browser via ONNX Runtime Web.

Efficient sub-pixel convolutional neural network for real-time single image super-resolution. Doubles or quadruples resolution without the compute cost of larger diffusion upscalers.

OpenAI's open-source automatic speech recognition model. Transcribes speech in 99 languages directly in the browser via Transformers.js — no audio ever leaves the device.

Meta's massively multilingual speech text-to-speech model. Provides AI-generated voice output alongside the browser's native SpeechSynthesis voices.

Google's on-device ML framework. Runs real-time video segmentation (SelfieSegmenter) for background removal and face detection for privacy masking — entirely client-side via WebAssembly.

Hugging Face's library for running transformer models directly in the browser via WebGPU and WebAssembly. Loads Whisper, MODNet, ESPCN, and MMS TTS on-demand with Cache API persistence so the second visit is instant.

Microsoft's cross-platform ML inference engine. Runs neural network models in the browser with WebGPU acceleration under the hood of Transformers.js.

Lightweight C/C++ inference engine for Stable Diffusion family models. Runs the FLUX 2 Klein 4B model on a single T4 GPU with minimal memory overhead.

The industry-standard open-source toolkit for video and audio processing. Handles all server-side video encoding, cropping, merging, resizing, speed changes, and format conversion.

FFmpeg compiled to WebAssembly for in-browser video compression and encoding without server uploads — your files never leave your device.

WebCodecs-based media toolkit for browser-side video conversion, thumbnail extraction, and muxing. Powers the video studio encoder, frame sampling, and MP4/WebM/MOV conversion.

Bitmap-to-vector tracing algorithm compiled to WebAssembly. Converts raster images into clean, scalable SVG paths.

Mozilla's JavaScript PDF rendering engine. Powers in-browser PDF preview, text extraction, and page-to-image conversion for every PDF tool that reads.

Pure-JavaScript PDF creation and manipulation library. Drives every PDF tool that writes — merging, splitting, rotating, reordering, and Markdown-to-PDF generation.

Hardware-accelerated 2D WebGL renderer. Powers the video studio compositor, real-time effect preview, and the slow-motion frame interpolator.

Library of GPU shader effects for PixiJS — blur, bloom, color matrix, displacement, and more. Drives the full effect stack in the video studio.

Fast, zero-dependency animated GIF encoder with color quantization. Produces the animated GIF export option from the video studio.

WebAssembly build of libheif for decoding HEIC/HEIF images from iPhones directly in the browser. Lets every image tool accept iOS camera output natively.

WebAssembly port of Google's woff2 font compression tool. Produces WOFF2 output for web-optimized fonts.

JavaScript font parser and writer supporting TrueType and OpenType formats. Powers the online font editor, fancy text generator, and font converter.

Data visualization library for dynamic, interactive graphics. Renders the mindmap tool's collapsible tree layout with smooth D3 transitions.

QR code generator with PNG and SVG output. Supports URL, WiFi, vCard, email, and other data types with customizable error correction.

XSS-safe HTML sanitizer by Cure53. Cleans all LLM-generated and user-supplied HTML before rendering — critical for the HTML tools that execute arbitrary markup.

Low-level Markdown compiler. Parses Markdown input for PDF generation and conversion tools, plus LLM chat outputs across the site.

Client-side ZIP archive reader and writer. Packages multi-file bulk-tool outputs into a single download without a server round-trip.

Production-grade animation library for React (formerly Framer Motion). Powers page transitions and UI micro-interactions across every tool page.

Open-source icon library (community fork of Feather). Supplies every icon in the navigation, tool cards, and in-tool controls.

Next-generation browser API for GPU-accelerated compute. Enables AI models (Gemma, Whisper, MODNet) to run at near-native speed in the browser on capable devices.

Binary instruction format for stack-based virtual machines. Enables near-native performance for FFmpeg WASM, Potrace, libheif, and every browser-side media library.

Browser-native speech recognition and synthesis. Powers live speech-to-text transcription and text-to-speech output with 200+ system voices.

Browser-native 2D drawing surface for image manipulation, text rendering, video frame processing, and format conversion. Used by every tool that composites or draws.

Browser-native media recording from canvas streams. Enables client-side video encoding for tools that produce new video output.

Browser-native audio processing graph for real-time audio manipulation, pitch preservation during speed changes, and encoding.

React framework with server-side rendering, file-based routing, and optimized bundling. The foundation of AICreate.

Declarative UI library for building component-based interfaces. All tool UIs are React components.

Utility-first CSS framework for rapid UI development with dark mode and responsive design built in.

Typed superset of JavaScript. The entire codebase is written in strict TypeScript for safer refactors.

Vite-native unit testing framework. Runs the library and hook test suite.

Microsoft's end-to-end browser automation framework. Drives cross-browser (Chrome + WebKit) tests against every deployed tool.

Pluggable JavaScript and TypeScript linter. Enforces code quality and catches SSR footguns before they ship.

High-performance reverse proxy. Terminates TLS, rate-limits traffic, caps request bodies, and serves static assets in production.