CloudFlare ImgBed

CloudFlare ImgBed is an open-source file hosting and image bed solution that supports both Docker and Serverless deployment. It allows you to unify multiple storage backends into a single management interface.

Supported Storage Channels:

• Telegram
• Discord
• Cloudflare R2
• S3 compatible storage
• Hugging Face
• WebDAV

Key Features:

• File management and directory organization
• Identity authentication and user management
• Content auditing
• RESTful API support
• WebDAV support
• Suitable for personal image hosting, website resource management, and lightweight file distribution.

The project supports an ecosystem of community-driven tools and integrations:

• Plugin Extensions: Browser extensions, integrations for Typecho, WordPress, and Obsidian, and OpenList drivers.
• Companion Applications: Desktop clients and bot tools.
• AI Agent Applications: Official project skills and related tools.
• Tutorials and Guides: Community-created videos and articles.

Developers can contribute to the ecosystem by submitting work via the Ecosystem Call for Contributions.

You can deploy the cloudflare-imgbed service using Docker Compose. The service maps the internal port 8080 to host port 7658 and persists data in a local ./data directory.

version: '3.8'

services:
  imgbed:
    image: marseventh/cloudflare-imgbed:latest
    ports:
      - "7658:8080"
    volumes:
      - ./data:/app/data
    restart: unless-stopped

For large files that exceed standard upload limits, use the chunked upload workflow by passing specific query parameters to the upload endpoint:

1. Initialize: Call the endpoint with ?initChunked=true to start a session.
2. Upload Chunks: Upload individual parts using ?chunked=true.
3. Merge: Once all parts are uploaded, call the endpoint with ?chunked=true&merge=true to combine the chunks into a single file.
4. Cleanup: If an upload is interrupted, use ?cleanup=true&uploadId={id}&totalChunks={count} to clean up orphaned chunks.

The Cloudflare Worker deployment entrypoint is automatically generated to map incoming requests to the appropriate business logic modules. To regenerate this file after making changes to the functions/ directory, run the following command:

node deploy/worker/generate-routes.js

Warning: Do not edit deploy/worker/index.js manually, as your changes will be overwritten by the generation command.

The project includes a native Node.js server implementation designed for Docker environments. This server uses Hono as the web framework to proxy Cloudflare Pages Functions requests. It replaces Cloudflare-specific services with local alternatives:

• Database: Uses SqliteD1 (SQLite) instead of Cloudflare D1.
• Storage: Uses LocalR2Storage (local file system) instead of Cloudflare R2.
• Functions: Proxies requests to the functions/ directory, simulating the Cloudflare Pages Functions runtime.

The server listens on the port specified by the PORT environment variable (defaults to 8080).

When running the native Node.js server, the following environment variables and behaviors are used:

• PORT: The port the server will listen on (default: 8080).
• Automatic Data Directory: The server automatically creates a data/ directory in the project root to store the SQLite database and R2 local storage files.
• Simulated Cloudflare API: The server injects a mock globalThis.caches object to prevent errors in code expecting the Cloudflare Cache API.
• IP Address Handling: The server attempts to extract the client's real IP using @hono/node-server/conninfo and injects it into the x-real-ip header to simulate Cloudflare's cf-connecting-ip behavior.

The Random Image API behavior is controlled via the othersConfig fetched from the environment. To enable or restrict the API, you must configure the following keys in your system configuration:

• randomImageAPI.enabled: Boolean. Set to true to allow access to the random endpoint.
• randomImageAPI.allowedDir: Comma-separated string of directory paths. The dir parameter in the request must match one of these paths or be a subdirectory of one.

If you want to upload a file from an existing web URL rather than sending raw bytes, use the external channel. This requires providing the url parameter in the form data.

Request Requirements:

• uploadChannel: external
• formdata: Must contain a field named url containing the source link.

The onRequest function serves as an API endpoint to fetch the latest Bing wallpapers. It fetches data from the Bing Image Archive and returns a JSON response containing an array of image objects.

Response Format: Returns a JSON object with the following structure:

• status (boolean): Indicates if the operation was successful.
• message (string): A status message (e.g., "操作成功").
• data (array): An array of image objects retrieved from Bing.