Privacy Policy

Introduction

Essence Scholar ("we," "our," or "us") is committed to protecting your privacy. This Privacy Policy explains how we collect, use, and safeguard your information when you use our research tools (browser extension, website, and the essencescholar-mcp connector for Claude) and our education platform (EssenceTeacher / EssenceStudy at study.essencescholar.com).

Claude / MCP Connector (essencescholar-mcp)

The essencescholar-mcp connector lets Claude search economics literature, run deep research sessions, manage your Essence Scholar paper library, and run or author Agent Studio workflows on your behalf. It operates on the same account, library, and backend as the web app — the practices described throughout this policy apply to it too. This section covers what's specific to the connector.

• Authentication: You supply your personal Essence Scholar API key when you connect the tool in Claude. Claude stores it locally (as a sensitive/masked value) and sends it as a bearer token with each request; we never see or store it anywhere the key wasn't already used for API access via the web app.
• What is sent: Only the arguments of the specific action you or Claude invoke — e.g. a search query, a paper's local file path (for reading), extracted text or PDF bytes, a research topic/description, or a workflow you're authoring. Nothing is sent unless a tool is explicitly called.
• No background access: The connector cannot read Claude's conversation history, memory, or any files beyond what a tool call explicitly names (e.g. a PDF path you point it to). It only acts when a tool is invoked, and only on the parameters passed to that call.
• Storage: Papers, research sessions, and workflows created through the connector are stored in your Essence Scholar library exactly like content added via the web app — same retention and deletion rights described in "Data Retention" and "Your Rights and Choices" below.
• Shared workflows: If you explicitly mark a workflow as shared to the Agent Studio marketplace, its steps and prompts (not your paper content) become visible to other users, same as sharing via the web app.

What We Collect

Information We Collect

• Account Information: Your email address, display name, and a hashed password (or your Google account identity if you sign in with Google). Credits and per-feature usage events are recorded against your account for billing.
• Academic Paper Content: PDF bytes and metadata (title, authors, DOI) sent to our backend for ingestion, OCR, and AI analysis, and stored in your personal library
• Content You Create: Chats, notebooks, research sessions, workflows, submission-check analyses, and researcher profiles you build are stored in your account until you delete them
• API Key: Your Essence Scholar API key is stored locally (browser chrome.storage for the extension, your MCP client for Claude) to authenticate requests to our backend
• Usage Analytics: The website uses Vercel Analytics (cookieless, aggregated page metrics), which loads only if you accept analytics cookies, and Cloudflare Turnstile for bot protection on login
• Product Analytics (first-party): For signed-in users we record a small number of structured product events — such as which onboarding option you chose and whether an analysis completed — against your account, so we can see where the product is confusing and fix it. No tracking cookie, no device identifier, no third-party analytics provider, and no free-form content from your documents or messages. These events are deleted when your account is deleted.

Information We Do NOT Collect

• Browsing history outside of pages where you explicitly trigger an import
• Health, financial, or payment information
• Location data or general web history
• Advertising identifiers of any kind — we run no ads and sell no data

How We Use Your Information

Core Functionality

• Paper Analysis: Process academic content to generate AI-powered analyses, reviews, and research outputs
• Your Library: Keep your papers, chats, and research artifacts available across sessions and devices
• Billing: Meter feature usage against your credit balance
• Email: Send account emails and, only if you subscribe, weekly literature newsletters (you can unsubscribe at any time)

AI Processing & Sub-Processors

AI features send the content being analysed (paper text, your prompts, and relevant context) to large-language-model providers. We use:

• Google Cloud Vertex AI (Gemini models) — our primary AI provider. Content is processed under Google Cloud's data-processing terms; Google Cloud does not use customer content submitted to Vertex AI to train its models.
• OpenRouter — our fallback provider, which routes requests to the selected model's operator (e.g. DeepSeek, OpenAI, Meta). Used automatically when the primary provider is unavailable, and for non-Gemini models.

Only the content needed for the specific analysis is sent; your account identity is not attached to model requests beyond what routing technically requires.

Other Service Providers

• Hosting: Backend servers in the European Union (Netherlands; standby capacity on Google Cloud Run, region europe-west4). Frontend hosted on Vercel. Traffic is proxied through Cloudflare.
• Database: Neon (PostgreSQL), hosted in the EU (Frankfurt). Encrypted backups are retained for 30 days.
• Email: Brevo processes transactional and newsletter email on our behalf (your email address and the newsletter content only).
• Search & Metadata: Literature features query academic services (Crossref, OpenAlex, Semantic Scholar, SSRN, arXiv, EconPapers) and, for web search, Serper. These receive search terms and paper identifiers — never your account identity.

No Advertising

• We do not use advertising networks
• We do not sell or share data with advertisers

Education Platform (EssenceTeacher / EssenceStudy)

Our education platform at study.essencescholar.com lets teachers create courses and assignments and students build AI research workflows and submit reports. It runs on the same infrastructure and AI providers described above (EU hosting, Google Vertex AI primary with OpenRouter fallback). Additional specifics:

• Accounts & roles: Sign-in is via Google (institutional accounts, e.g. University of Groningen @rug.nl / @student.rug.nl) or email/password. We store your email, name, and role (student, teacher, or admin).
• Course & enrolment data: Course membership, teacher assignments, and the courses you are enrolled in.
• Coursework you create: Uploaded documents (PDF/CSV/URL), the workflow canvases you build, and the reports (submissions) you generate — stored against your account and the assignment.
• Assessment data: When a teacher enables the Assessment Module, AI-drafted grades, teacher-edited grades, rubric scores, and written feedback are stored on your submission. Students see grades only after a teacher publishes them.
• Visibility: Teachers and course admins can view the coursework and submissions of students enrolled in their courses; students see only their own work and published grades. Administrators can manage accounts and, for support, impersonate a user — such actions are limited to platform operation.
• AI grading & analysis: Report generation and AI grade drafts send the relevant document and workflow content to our AI providers, as described in “AI Processing & Sub-Processors” above.
• Minimum age: The education platform is intended for university students and staff; it is not directed at children under 16.

Your Rights and Choices

Data Control

• Access: View your stored preferences and settings
• Delete: Clear all stored data at any time
• Opt-out: Disable the extension to stop data collection
• Export: Request a copy of your stored data

Browser Permissions

The extension requires the following permissions:

• activeTab: To read the current tab's URL and detect PDFs
• scripting: To inject a content script that extracts PDF bytes from the page
• tabs: To get active tab information for triggering imports
• storage: To save your API key and extension preferences locally
• webRequest: To detect PDFs at the network level by inspecting Content-Type headers
• webNavigation: To re-check PDF status when you navigate to a new URL
• host_permissions (<all_urls>): Academic PDFs are hosted on any domain — publishers, preprint servers, institutional repositories — so universal access is required

Data Retention

Retention Period

• Your Library & Creations: Papers, chats, analyses, and workflows are kept until you delete them or your account
• Usage Records: Credit/usage events are retained for billing integrity
• Backups: Encrypted database backups are retained for 30 days, then deleted
• Local Preferences: Extension settings stay on your device until you clear them

Deletion

• Manual: Delete individual papers or artifacts from the web app at any time
• Account: Contact us to delete your account and all associated data
• Extension Removal: All locally stored data is deleted when the extension is uninstalled

Children's Privacy

Our extension is designed for academic research and is not intended for children under 13. We do not knowingly collect personal information from children under 13.

International Users

Essence Scholar is operated from the Netherlands, and our servers and database are located in the European Union. Content sent to AI providers for analysis may be processed outside the EU (e.g. by Google Cloud or a model operator routed via OpenRouter) under those providers' data-processing agreements and applicable transfer safeguards.

Changes to This Policy

We may update this Privacy Policy from time to time. We will notify you of any changes by:

• Posting the new Privacy Policy on our website
• Updating the "Last updated" date
• Notifying users through the extension if significant changes occur

Legal Basis

This Privacy Policy is provided in compliance with:

• Chrome Web Store Developer Program Policies
• General Data Protection Regulation (GDPR)
• California Consumer Privacy Act (CCPA)
• Other applicable privacy laws

Data Protection Officer

For privacy-related inquiries, please contact: