Scrolljar Backend

Scrolljar is a pastebin-style backend service where users create Jars (collections) containing multiple Scrolls (📜 pastes). This repository contains the API server, storage integration, and cleanup tooling.

Core Concepts

• Jar: A collection of scrolls. Can be public or private and optionally password-protected.
• Scroll: A single paste belonging to a jar.
• User: Authenticated users can manage jars and scrolls.
• Anonymous users: Can create jars without signing in.

Docs

Swagger Docs

Data Flow

View a Diagram

Writing (Create & Upload)

1. Client sends jar/scroll metadata (name, password, expiry, etc.) to a create endpoint.
2. API:
   • Generates unique Base62 IDs for jars and scrolls
   • Stores metadata in PostgreSQL
   • Issues a short-lived upload token
3. Client uploads scroll content using PUT /upload with the upload token.
4. API streams the request body directly to S3:
   • No in-memory storage
   • Enforces size and encoding restrictions
   • Returns immediate errors on validation failure

Reading (Fetch)

1. Client sends a GET request with the jar or scroll ID.
2. API fetches metadata from the database.
3. If password-protected, the password is provided via the X-Paste-Password header and verified.
4. API returns a presigned S3 URL for direct content access.

Authentication

• Bearer token–based authentication
• Required for:
  • Deleting jars and scrolls
  • Updating scrolls
  • Accessing user and user-owned resources

Rate Limiting

• Global rate limit applied to all requests
• IP-based limits:
  • Strict: user registration, activation
  • Medium: create, update, delete operations
  • General: read operations (jar / scroll)