Writademic
Sign in to your account
Sign In
Create Account
Last Updated: June 2026
Ethics & data-protection framework
Writademic is designed to comply with major academic-AI ethics frameworks (COPE, ICMJE, Russell Group AI principles, and equivalent funder guidance) and with applicable data-protection law in your jurisdiction (GDPR in the EU/UK, CCPA in California, KVKK in Türkiye, LGPD in Brazil, and equivalents). The operational guarantees described below apply uniformly to every user, regardless of jurisdiction.
Writademic is built on the principle that AI must not assist an author whose voice it has not learned. Before any generative or editorial endpoint will run for you, a personal style baseline of at least 5,000 words of your own academic writing must be analyzed. This is not a marketing choice — it is the technical guarantee that AI suggestions are shaped to your voice, not a generic LLM register. This requirement directly implements the "meaningful human oversight" principle that is central to every contemporary academic-AI ethics framework.
We know academic researchers handle pre-published, sensitive, and proprietary data. Your manuscripts, inputs, and outputs are never used to train our models or any third-party LLM.
Writademic delegates language-model inference to Anthropic, Inc. (headquartered in San Francisco, California, United States), the operator of the Claude family of large language models. When you submit text for any AI operation, that text is transmitted to Anthropic's API for the duration of the call and a response is returned to you. The specific Claude model identifier used at the time of each call (e.g. claude-opus-4-7, claude-haiku-4-5-20251001) is recorded in your AI Usage Statement so that the disclosure you produce for a journal, supervisor, or ethics board names the exact tool used.
Each outbound request carries a private-tier header (X-Writademic-Privacy: no-retention; no-training) and is sent under Anthropic's commercial Zero Data Retention (ZDR) configuration, which means: (i) Anthropic does not retain your prompts or completions on its infrastructure after the API response is returned; (ii) your data is not used to train Anthropic's models; (iii) no human review of your content occurs.
If your institution, funder, or research protocol forbids transmitting sensitive material (for example unpublished clinical trial data, identifiable patient records, classified or trade-secret information) to a third-party AI provider located outside Türkiye, do not paste that content into Writademic.
Because Anthropic's inference servers are located in the United States, any personal data contained in text you submit to Writademic is technically subject to a cross-border transfer under whichever data-protection regime applies to you — for example, GDPR Articles 44–49 (EU/EEA/UK), CCPA §1798.130 (California), KVKK Madde 9 (Türkiye), or LGPD Art. 33 (Brazil).
By registering for and using Writademic, you provide explicit, informed, and specific consent for the limited transfer of any personal data contained in your submitted text to Anthropic, Inc. for the sole purpose of producing the AI output you requested. This consent is informed (you are reading this section), specific (limited to AI inference for the requested operation), and freely given (you may refuse by not using AI tools — your account remains usable for style analysis and the AI Usage Statement export).
To minimize the volume of personal data crossing the border, you are strongly advised to remove or pseudonymize personal data (names of human subjects, national identifiers, patient identifiers, contact details of third parties, identifiable interview transcripts) before pasting text into Writademic. You remain the data controller for any personal data inside your submissions and are responsible for that classification under your applicable law.
Writademic itself is the data controller for the account metadata it stores (your name, email, hashed password, style profile, operation history). Anthropic acts as a data processor for the duration of each API call only, under Anthropic's published Data Processing Addendum (DPA). A copy of Anthropic's DPA is available at anthropic.com/legal/commercial-terms.
/api/ai-usage-statement (JSON / Markdown / plain text / signed PDF). The PDF includes the platform name, the exact LLM model identifier and version used, the date range, the operations performed, and a public verification code. Use it to disclose AI assistance to a journal, supervisor, or ethics board.The following inputs are not appropriate for any third-party AI tool, including Writademic:
Writademic cannot detect every category automatically. The final responsibility for filtering personal or confidential content before paste remains with you under your applicable data-protection law.
Major academic-AI ethics frameworks and funder policies (NIH, NSF, ERC, and equivalents) prohibit the use of generative AI in the peer-review workflow because of confidentiality and integrity concerns. Writademic does not offer reviewer tools — endpoints such as /api/reviewer-response are for applicants drafting replies to reviewer comments, not for reviewers evaluating proposals. If you are acting as a reviewer or panelist, do not use Writademic on materials you are evaluating.
Writademic does not and will never sell your personal data, institutional affiliations, or uploaded manuscripts to data brokers, advertising networks, or AI companies.
You may export all your account data (style profile, operation history, feedback) or delete your account at any time from the dashboard. Deletion removes your account row, your uploaded reference texts, and your style profile from our database. Under your applicable data-protection law you also have rights to access, rectification, erasure, restriction, portability, and objection (GDPR Art. 15–22 / CCPA §1798.100–1798.140 / KVKK Madde 11 / LGPD Art. 18, as applicable). Direct any such request to the contact below.
For questions about this policy or to exercise your data-protection rights, contact our privacy team at info@writademic.com.