moar: A Chrome Extension That Saves 95% of Tokens Through Local Document Processing

Overview

When using ChatGPT, Claude, or Gemini to process documents, do you frequently hit "file too large" limits? Or get poor AI responses because of messy document formatting? moar is a free Chrome extension designed to solve the document preprocessing challenge of the AI era — it extracts the true structure from 9 formats including PDF and DOCX, outputs clean Markdown or CSV, saves up to 95% of Token consumption, and completes all processing locally on your device.

Core Features of moar

Intelligent Document Structure Extraction

moar's core value lies in "structure extraction" rather than simple format conversion. It recognizes semantic structures within documents — heading hierarchies, tables, lists — and transforms them into Markdown or CSV formats that AI understands best.

Markdown is a lightweight markup language that uses concise symbols like # for heading levels, ** for bold, and - for list items to describe document structure. Since LLM training corpora contain vast amounts of Markdown text from platforms like GitHub and Stack Overflow, models have exceptionally strong comprehension of this format. Compared to disordered text extracted from raw PDFs or tag-heavy HTML, well-structured Markdown helps AI more accurately identify paragraph relationships, table data, and logical hierarchies, significantly improving response quality. AI tools receive semantically optimized structured content instead of chaotic raw text.

Supported input formats cover the main everyday office scenarios:

Documents: PDF, DOCX, TXT, MD
Presentations: PPTX
Data: XLSX, CSV, JSON
Web: HTML

Single files up to 50MB are supported, covering the vast majority of everyday documents.

How 95% Token Savings Works

This is moar's most compelling metric. To understand it, you first need to know what Tokens are: Tokens are the basic unit of measurement for LLM text processing. In English, one Token corresponds to roughly 4 characters or 0.75 words; in Chinese, one character typically equals 1-2 Tokens. Major AI services charge or set limits based on Tokens — GPT-4 has a context window of 128K Tokens, and Claude 3.5 Sonnet supports 200K Tokens.

When we paste document content directly into an AI conversation, PDF format metadata, HTML tags, redundant whitespace, repeated information, and other "useless content" consume precious Token quota in large quantities — driving up costs while squeezing the processing space for actually useful content. Through intelligent compression and structural processing, moar reduces Token consumption by up to 95% with zero semantic loss.

For developers and heavy AI users on paid APIs, this means significant monthly cost savings. For users on free tiers, it means the same quota can process substantially more content.

100% Local Processing for Privacy Protection

moar adopts the strictest privacy-by-design approach: all document processing happens entirely on your local device, and files never leave your computer. This is particularly important for handling sensitive business documents, legal contracts, financial statements, and similar scenarios.

This "zero-upload" architecture relies on the maturation of WebAssembly (WASM) technology. WASM allows high-performance document parsing libraries written in C/C++ or Rust (such as PDF.js, mammoth.js, etc.) to be compiled into binary formats that run within the browser sandbox, completing complex file parsing tasks without server-side support. This architecture not only achieves near-native processing speed and full offline capability, but also inherently satisfies data compliance requirements like GDPR, while avoiding the operational costs of cloud processing services — making the free model sustainable.

Users don't need to worry about document content being stored or analyzed by third-party servers, and enterprise compliance scenarios can use it with confidence.

Who Should Use moar

Target User Profiles

Heavy AI users: Knowledge workers who need to feed large volumes of documents to AI for analysis daily
Developers: Technical professionals who call AI services via API and are sensitive to Token costs
Enterprise users: Business professionals who need to process sensitive documents while leveraging AI for efficiency
Researchers: Academic workers who need to batch-process papers, reports, and other long documents

Typical Workflow

Install the moar extension in Chrome
Upload the document you need to process (supports 9 formats, single file up to 50MB)
moar extracts the document structure locally and generates optimized Markdown or CSV
Paste the processed content directly into ChatGPT, Claude, or Gemini

The entire process requires no account registration, no online upload, and conversion completes in seconds.

Market Positioning and Comparison with Similar Tools

moar is positioned in the "preprocessing" stage of the AI toolchain, filling the gap between raw documents and AI input. The context for this demand: as enterprise AI applications become widespread, RAG (Retrieval-Augmented Generation) and document intelligence have become hot topics, creating strong demand for document preprocessing tools. Professional document parsing services like Adobe PDF Extract API, AWS Textract, and Microsoft Azure Document Intelligence are all paid cloud services primarily targeting enterprise developers. For individual users and small teams, this stage has long lacked lightweight tools. Currently, similar local document preprocessing tools are rare in the market — most users still manually copy-paste or directly upload raw files to AI.

As a free Chrome extension, moar has an extremely low barrier to entry — just install and use. However, its long-term business model remains unclear — whether it will introduce paid premium features, enterprise versions, or processing limits is worth watching.

Conclusion

moar solves a real and widespread pain point: getting documents into AI workflows in the most efficient and secure way possible. 95% Token savings, privacy protection through local processing, and broad support for 9 formats — these three core selling points form a clear product value proposition. If you frequently use AI to process various documents, moar is a Chrome productivity extension worth installing.

Key Takeaways

moar is a Chrome extension that extracts structured content from 9 document formats, outputting AI-friendly Markdown or CSV
Saves up to 95% of Token consumption while maintaining zero semantic loss
All document processing is 100% local — files are never uploaded to external servers
Supports PDF, DOCX, PPTX, XLSX, CSV, TXT, MD, JSON, and HTML formats, with a single file maximum of 50MB
Free to use, compatible with mainstream AI tools including ChatGPT, Claude, and Gemini