Skip to content
curiositi

Curiositi Documentation

Curiositi is your digital brain — an AI-powered application that helps you store, organize, and effortlessly find any file using semantic search.

What is Curiositi?

Section titled “What is Curiositi?”

Curiositi transforms how you manage digital content. Instead of remembering exact filenames or folder locations, simply describe what you’re looking for in natural language. The AI understands meaning and context to surface the right files instantly.

Smart Storage

Upload PDFs, images, text files, and more. Files are stored in S3-compatible storage and processed automatically.

Semantic Search

Find files by meaning, not just keywords. Ask “quarterly sales report” and find relevant documents even if they don’t contain those exact words.

AI Processing

Every file is automatically processed to extract content, chunk it, and create searchable vector embeddings.

Fast and Scalable

Built on PostgreSQL with pgvector, Upstash QStash for async processing, and Bun for performance.

Spaces

Organize files in hierarchical spaces. Create nested structures that match how you think.

Workspaces

Collaborate with teams using multi-tenant workspace support. All content is scoped to your workspace.

Quick Start

Section titled “Quick Start”

Get up and running in minutes:

Terminal window
# Clone the repository
git clone https://github.com/macintushar/curiositi.git
cd curiositi


# Install dependencies
bun install


# Set up environment variables
cp .env.example .env


# Start development servers
bun run dev

Visit the Getting Started guide for detailed setup instructions.

How It Works

Section titled “How It Works”
  1. Upload — Drag and drop any supported file into Curiositi
  2. Process — The worker extracts content and generates embeddings in the background
  3. Organize — Add files to spaces, create nested hierarchies
  4. Search — Find anything using natural language queries

Supported File Types

Section titled “Supported File Types”
CategoryFormats
DocumentsPDF
TextPlain text (.txt), Markdown (.md), CSV
DataJSON, XML
WebHTML
ImagesJPEG, PNG, WebP, GIF (with AI-generated descriptions)

Maximum file size: 50 MB

Architecture Overview

Section titled “Architecture Overview”

Curiositi is built as a monorepo with Turborepo and Bun:

  • Platform — React 19 web app with TanStack Start, tRPC v11, and Tailwind CSS v4
  • Worker — Hono server on Bun for AI-powered file processing
  • Database — PostgreSQL with Drizzle ORM and pgvector for embeddings
  • Storage — S3-compatible object storage for files
  • Queue — Upstash QStash for background job processing

Learn more in the Architecture section.

Next Steps

Section titled “Next Steps”