Collaborate. Create. Conquer.
PageSmith is a hyper-fast, real-time collaborative workspace built for technical teams, developers, and creators. It bridges the gap between distraction-free Markdown editors, rich visual word processors, and AI-powered assistants—all wrapped in a striking, high-contrast Neo-Brutalist design.
Modern teams suffer from severe context switching.
- You write documentation in Google Docs or Notion.
- You write code in VS Code.
- You brainstorm and debug using ChatGPT or Gemini in a separate browser tab.
- You share and convert files through various third-party tools.
PageSmith solves this fragmentation. It provides a single, unified "Canvas" where you can type at the speed of thought using Markdown, visually align and format text, drop in heavy PDF documents to extract text, and ask an AI assistant to generate code—all while your teammates watch your cursor move in real-time.
Most collaborative text editors force you to choose a side:
- WYSIWYG (Notion/Google Docs): Great for visual layout, but clicking through menus to format code blocks breaks a developer's flow.
- Pure Markdown (HackMD/Obsidian): Great for speed, but terrible if you just want to center a title or change a font family.
PageSmith is the bridge. It gives you the speed of slash commands and Markdown, the visual controls of a rich-text editor, the collaborative power of multiplayer syncing, and the intelligence of an AI IDE—paired with an unapologetically bold UI.
Write in pure Markdown, but use the formatting toolbar to apply alignments (left, center, right, justify), custom font families, and highlighters. We achieve this by seamlessly parsing injected HTML alongside standard Markdown.
Keep your hands on the keyboard. Type /js, /python, /table, /math, or /todo to instantly scaffold complex structures, equations, and code blocks with full syntax highlighting.
Built on WebSockets, changes appear instantly across all connected clients. See exactly what your team is typing, debugging, or formatting in real-time.
A built-in, slide-out panel powered by Google Gemini. Ask the AI to write code, summarize text, or brainstorm ideas directly beside your document, then copy-paste the results instantly into your Canvas.
Drop .txt, .md, code files (.ts, .py, etc.), or even .pdf files directly into the editor. PageSmith parses PDF documents locally in the browser and injects the text directly into your workspace.
Instantly compile your beautiful Canvas—complete with themes, code blocks, and math equations—into a clean, printable PDF with a single click.
A bold, high-contrast user interface with hard shadows, sharp corners, and carefully curated color palettes (Nord, Slate, Ocean, Light/Dark).
Built-in dashboard management for Owned and Shared canvases, secured by Clerk Authentication, with tiered subscriptions integrated seamlessly via Razorpay.
PageSmith integrates a persistent AI Assistant directly into the workflow context.
- Flow State Preservation: Instead of switching to a different tab to consult an LLM, the assistant lives in the same window.
- Technical Implementation: The frontend securely communicates with a Next.js API route (
/api/gemini), which acts as a proxy to Google's Gemini models. This ensures API keys are kept secure on the server while delivering fast, contextual responses to the client.
PageSmith employs a modern, decoupled architecture designed for scale, speed, and safety:
- Frontend Rendering Magic: The
Editorcomponent usesreact-markdownpaired withrehype-rawandrehype-katex. This is the secret sauce that allows PageSmith to render standard Markdown, LaTeX math equations, and HTML<div>tags simultaneously, enabling MS Word-style alignments inside a Markdown file. - Debounced Real-Time Engine:
A custom
useSockethook manages the WebSocket connection. When a user types, the input is throttled using auseDebouncehook to prevent overwhelming the server. The payload is emitted via acanvas-updateevent, which the backend broadcasts to all users in that specific Canvas "Room". - Client-Side File Processing:
File parsing happens entirely on the client side using the browser's
FileReaderAPI andpdfjs-distweb workers. This ensures zero latency and absolute privacy—your sensitive PDFs and code files are parsed locally, not uploaded to a server to be read. - Secure Payment Handshake: The Razorpay workflow uses a 3-step secure handshake: The frontend requests an order ID via an API route -> The user pays via the Razorpay modal -> The frontend sends the signature to the backend for cryptographic verification before upgrading the database record.
- Frontend: React, Next.js (App Router), Tailwind CSS, Framer Motion
- Backend & Real-time: Node.js, Express, Socket.io
- Database: Prisma ORM
- Markdown Parsing:
react-markdown,remark-gfm,remark-math,rehype-raw,rehype-katex,react-syntax-highlighter - AI: Google Gemini API
- Auth: Clerk
- Payments: Razorpay
- Icons: Lucide React
- Node.js (v18+)
- API Keys for: Clerk, Razorpay, and Google Gemini.
- Clone the repository:
git clone [https://github.com/yourusername/pagesmith.git](https://github.com/yourusername/pagesmith.git) cd pagesmith