Bibliography to Zotero Converter

Step 1: Choose Input Source

URL (Web Scraping) Fetch bibliography from Omeka website Local File Upload .txt, .docx, or .pdf file Paste Text Copy and paste bibliography entries

Source URL:

Use CORS proxy (recommended)

Automatically adds a CORS proxy to bypass browser security restrictions. Uncheck only if you're using a browser extension or local proxy. Learn more about CORS →

Choose File:

Choose file...

Supported formats: .txt (plain text), .docx (Word document), .pdf (PDF document)

Bibliography Text:

Tip: Separate entries with blank lines (triple newlines work best, like in the source files). This matches the format used by Omeka bibliography pages.

Step 2: Output Format

CSL-JSON Recommended for Zotero (native format) RIS Alternative citation format

Step 3: AI Model Configuration

Configure OpenRouter AI to parse bibliographic entries into structured Zotero-compatible format.

Model:

⚠️ Free Model Data Retention: Free models (marked with "Free") may retain your input data for training purposes. For sensitive bibliographic data, consider using paid models like Claude 3.5 Sonnet or GPT-4o which have data privacy guarantees.

Popular models for bibliography parsing. Claude and GPT models work best. Browse all models →

Custom Model Name:

Enter the OpenRouter model identifier (format: provider/model-name). Find models at openrouter.ai/models

OpenRouter API Key:

Remember API key in browser

⚠️ Security Warning: Only check "Remember API key" if you're on a personal computer. Do NOT use this option on shared or public computers, as your API key will be stored in the browser's localStorage and could be accessed by others.

Get your API key at openrouter.ai/keys.

Batch Size:

Number of entries to process per API call (default: 25)

Step 4: Output Options

Test Mode (Process only 10 entries)

💡 What is Test Mode?
When enabled, only the first 10 bibliography entries will be processed. This is useful for:

Quick testing without using many API credits
Verifying the output format before full processing
Checking if your API key and model work correctly

Uncheck to process all entries (recommended after testing).

Output Filename:

The filename for the downloaded output file

Processing...

Initializing...

Results

Preview (First 5 Entries)

CORS (Cross-Origin Resource Sharing) is a browser security feature that prevents websites from accessing content on other domains. This is why direct web scraping can be blocked.

Solution 1: CORS Proxy (Default, Recommended)

This app automatically uses corsproxy.io to bypass CORS restrictions. The proxy acts as an intermediary, fetching the page on your behalf.

How it works:
Your URL: https://example.com/page
Becomes: https://corsproxy.io/?https://example.com/page

Other proxy options:
• https://api.allorigins.win/raw?url=
• https://cors-anywhere.herokuapp.com/ (may require temporary access)

Solution 2: Browser Extension

Install a CORS-disabling extension to allow direct access:

Chrome/Edge: "Allow CORS: Access-Control-Allow-Origin"
Firefox: "CORS Everywhere"

⚠️ Security Warning: Remember to disable the extension after use, as it reduces browser security protections.

If using an extension, uncheck "Use CORS proxy" above.

Solution 3: Manual Download & Upload

If automated scraping doesn't work:

Visit the webpage in your browser
Save the page (Ctrl+S / Cmd+S) or copy the bibliography text
Create a .txt file with the content
Use the "Local File" or "Paste Text" input option

Solution 4: Local Proxy Server (Advanced)

Run your own CORS proxy locally:

                            # Install cors-anywhere

npm install -g cors-anywhere

# Run proxy on port 8080

cors-anywhere

# Update proxy URL in the app's JavaScript

This web application converts bibliographic entries from URLs, local files, or pasted text into Zotero-importable formats (CSL-JSON or RIS). It uses AI to intelligently parse and structure bibliographic data for seamless import into reference managers.

Features

Multiple Input Methods: Web scraping with automatic CORS proxy, file upload (.txt/.docx/.pdf), or direct text paste
AI-Powered Parsing: Uses OpenRouter API to access multiple AI models (Claude, GPT, Gemini, etc.)
Smart Processing: Automatic entry splitting, author ditto mark expansion, and format detection
Flexible Output: CSL-JSON (recommended) or RIS format for Zotero import
Batch Processing: Handles large bibliographies with configurable batch sizes and retry logic
Error Handling: Failed entries are tracked and can be downloaded for manual review
Convenience Features: API key storage, test mode (10 entries), progress tracking

Getting Started

Get an API key from OpenRouter
Choose your input source (URL, file, or paste text)
Select an AI model (Claude Sonnet 4.5 or GPT-4o recommended)
Enter your API key (optionally save it for future use)
Enable Test Mode for initial testing (processes 10 entries)
Click "Process Bibliography" and wait for results
Download your CSL-JSON or RIS file and import into Zotero

Tips

Use Test Mode first to verify everything works before processing large bibliographies
GPT OSS 120B (default) is very cheap - other models available if you prefer different options
Batch size of 25 is optimal for most models (lower if you get timeouts)
CORS proxy is enabled by default - uncheck only if using a browser extension
Save your API key only on personal computers, never on shared devices

Project Info

This is a web version of the biblio-zotero Python script. Built with vanilla JavaScript, no backend required - runs entirely in your browser.