How does it work & what does it do?
Part 1: The Workflow — Step by Step
The Secure Academic Proofreader is designed to provide reliable linguistic and formatting analysis without compromising the privacy of your unpublished research. Here is exactly how to use the tool from start to finish.
Local Reading & Preparation
When you drag and drop your .docx file onto the application, it is not uploaded to our servers. Instead, the application uses your browser's local processing power to read the text.
Configuration & Pre-flight
Once the document is locally processed, the configuration wizard opens. Here you can select the modules you wish to run: Grammar & Spelling, Academic Tone, and the premium Citation Audit (supporting APA, Chicago, and MLA). You also set the primary language of the text.
The Live Cost Calculator dynamically displays the required credits based on the length of your text (in 900-character blocks) and the chosen modules. You can enter your anonymous Credit Pack Code here to activate the analysis.
Secure AI Analysis
By clicking "Launch Proofreader", the tokenized text is transmitted to our secure European cloud infrastructure. A progress window will appear, detailing the stages of the analysis.
Our multi-agent AI system concurrently checks your document for syntactical errors, phrasing, and citation formatting. As soon as the analysis is completed, the uploaded source data is scheduled for immediate deletion from our servers.
Review & Export
The results are displayed in a dual-pane table where you can review each issue. The application shows the original sentence, the suggested correction, and an academic reasoning for the change. You can accept or reject changes individually or in bulk.
- Download Corrected Document: Generates a clean HTML file containing all accepted fixes.
- Download Errata Report: Exports a full list of issues and reasoning for documentation.
- Save Progress (.sau): Saves your current workflow state securely.
Part 2: Data Security & Privacy Architecture
The promise of privacy is cheap; verifiable architecture is what matters. Secure Academic Studio was built from the ground up to handle highly sensitive intellectual property, unpublished research, and confidential academic manuscripts. Below is a transparent breakdown of the security measures protecting your documents.
Zero File Retention (Five-Layer Deletion Guarantee)
Unlike standard proofreading tools that store your documents indefinitely to build user profiles, our system employs a strict Zero Retention architecture. The text you process is only stored for the exact duration of the AI analysis, and its deletion is guaranteed by five independent layers — two enforced at the storage-infrastructure level, three enforced in our application code:
- Storage-Level Hardening: Soft delete and object versioning are disabled on the storage bucket itself, so any deletion issued by the layers below is truly final, with no recoverable copy left behind.
- On-Demand Erasure: The moment the analysis is complete and your browser downloads the results, the application actively instructs the server to immediately destroy the source text object.
- Process Finality: The backend process code includes an inescapable 'finally' block that permanently purges the data from the cloud, whether the analysis succeeds or fails due to an error.
- The Orphan Sweeper: As a fail-safe against network outages, an automated job sweeps our secure cloud environment every hour, deleting any temporary file older than 3 hours — in the rare worst case, up to roughly 4 hours, since a file must first cross the 3-hour mark and then wait for the next scheduled sweep.
- Cloud Provider Lifecycle Rule: As an absolute last resort, entirely independent of our own application code, a native cloud-storage lifecycle rule deletes any remaining object older than 1 day.
Client-Side Encryption (AES-256 .SAU Files)
Proofreading a 100,000-character manuscript takes time, and you may want to pause your work. Instead of saving your progress to a centralized database linked to an account, you can export your session as a .sau file.
If you choose to protect this file with a password, your browser's native Web Crypto API encrypts the entire dataset using military-grade AES-256-GCM encryption. The password never leaves your device, and the key is never transmitted to our servers. Your progress is stored locally on your hard drive, completely impenetrable without the password.
100% EU Server & AI Jurisdiction
Many "EU-based" tools silently route API requests to AI engines located in the United States. Secure Academic Studio operates a hardcoded EU-only AI routing protocol.
Our backend servers are hosted by Hetzner Online GmbH in Nuremberg, Germany. Our AI infrastructure is hardcoded at the code level to the EU region, and this is reinforced by an organization-wide Google Cloud policy that blocks any resource from being provisioned outside EU locations — together ensuring that all Vertex AI requests are processed exclusively within European data centers. We comply strictly with GDPR requirements, and your data is never used to train or fine-tune public AI models.
Structural Tokenization
Before your document is ever transmitted, the local Mammoth.js engine in your browser strips out metadata (such as tracked-changes author names and internal footnote back-reference links generated during conversion). Furthermore, structural HTML tags — headings, tables, and other formatting elements — are swapped with anonymous placeholders (e.g., [__T_1__]); italics, bold, and other emphasis tags are deliberately left untouched so the AI can still correctly interpret emphasis in your writing. The cloud AI only ever receives the raw textual content mixed with these meaningless tokens, preventing the leakage of your document's underlying structural integrity.
Complete Anonymity (No Accounts)
The most secure personal data is the data that is never collected. Secure Academic Studio operates entirely without user accounts. There is no registration, no email address collection, and no password database to be breached.
Access is managed via anonymous Credit Pack Codes tied only to a credit balance — never to your name, email, or payment details. Even our free-trial abuse prevention avoids storing personal data: instead of your raw IP address, we store a one-way, salted cryptographic hash whose input also includes the current date, so it cannot be used for long-term tracking across days. You remain completely anonymous.