Privacy Notice

Collect less.
Publish only research.

Effective 16 July 2026

Data minimisation by design.

Research Idea avoids public user profiles. Public reading needs no account. Registry records exclude raw sign-in emails and names. Completed research is connected through a public scholarly link, without uploading a copy to the Site. The authentication provider necessarily retains account information for passwordless access. Read this Notice with the Terms of Use.

01

Who is responsible

BDF Partners (Pty) Ltd, registration number 2025/857305/07, is a company registered in the Republic of South Africa and trading as Research Idea (formerly Research Idea Registry). It is the responsible party for personal information processed through the Site. Privacy and data-rights enquiries may be sent through the operator contact page.

02

Information we process

When you sign in, Supabase processes your email address to send and verify a one-time authentication code. It supplies the Site with a stable account identifier. The registry transforms that identifier into a protected, site-specific contributor code. Registry, usage and ownership records store this code while excluding the email and profile name. Readers cannot use the code to discover your account.

The Site also processes the research material you choose to submit: ideas, limited source excerpts, citations, scholarly annotations, research-status signals and public research links. Paper citations and related-research metadata necessarily contain publicly available author names. The terms-acceptance cookie contains only the accepted version, without a user identifier.

If you write to the Operator through the contact page, the Site stores the category you chose, the message you wrote and the reply address you supply, so that the enquiry can be answered and, where it concerns a payment, privacy request or legal complaint, so that a record of how it was handled exists. Contact messages are kept only until the matter is resolved and any applicable record-keeping period has passed. Supplying a reply address is optional, but the Operator cannot respond without one.

Do not submit participant-level data, names of private individuals, identity numbers, contact details, health records, confidential institutional information, embargoed work or other sensitive information. Obvious contact and identity numbers are blocked before text is sent to the AI service.

03

What becomes public

Published research briefs, pseudonymous contributor codes, scholarly annotations, provenance, source citations, research-status signals and connected research links are public. Research Idea does not intentionally publish or retain your sign-in email or profile name.

A public scholarly record may remain available for attribution and registry integrity even if you stop using the Site, subject to lawful correction, objection and deletion rights.

04

Why we use information

Information is used to authenticate contributors; create, classify and preserve public records; operate automated admission and source checks; prevent abuse; enforce limits and the Terms; connect research outputs; maintain attribution and version history; monitor the fairness and reliability of the published admission standard; secure and improve the service; respond to rights or legal requests; and comply with law.

Depending on the context, processing is based on your consent, performance of the Site agreement, the Operator’s legitimate interests in running a trustworthy scholarly registry, and applicable legal obligations.

05

Authentication, AI, scholarly and payment services

Supabase provides passwordless email-code sign-in, Resend delivers authentication emails and forwards contact-form messages to the Operator’s mailbox (hosted by Microsoft 365), and Cloudflare Turnstile checks whether an authentication request appears human. PayFast processes payment details for credit purchases and voluntary support. These providers process information under their own terms and privacy practices.

Natural-language searches and research text submitted for automated review are sent to Anthropic’s Claude service, and to Google Gemini as a backup when Claude is unavailable. Each provider handles inputs and outputs under its own terms for the service tier the Site uses; where those terms permit product improvement or human review, submitted material may be used in that way. The Site first rejects obvious email addresses, phone numbers and identity or passport numbers, but automated detection cannot identify every form of personal or confidential information.

Academic search terms and bibliographic details may be sent to Crossref, Semantic Scholar or OpenAlex to verify sources and find related work. The hosting provider handles requests needed to deliver and secure the Site.

06

International processing

External service providers may process information outside South Africa. The Operator reviews provider arrangements and applicable contractual or legal safeguards when providers or processing practices change.

07

Retention

Accepted AI review packages expire after 24 hours and are deleted after use; the contributor’s rough original wording is not retained with the published idea. Weekly AI-usage records and paid-credit ledgers contain the protected contributor code and the information needed to calculate balances, expiry and resets. Daily search, checkout and contact fair-use counters use protected account or network identifiers and are automatically removed after approximately 45 days. Contact-form messages are retained only until the enquiry is resolved and any applicable record-keeping period has passed. Payment records retain the transaction details described below. The Site also retains anonymous aggregate review totals by date, broad academic field, source type, outcome, failed gates and manipulation flag. These totals contain no contributor code, account identifier or raw draft and cannot be used to reconstruct a submission. Published scholarly records may be retained longer to preserve provenance and registry integrity. Source excerpts are used for review and are not intentionally published in the public record.

08

Security and deliberate limits

Research Idea uses passwordless authentication, human-verification controls, authenticated write actions, server-side enforcement and a secret-key transformation that replaces the provider account identifier with a site-specific contributor code. The registry database does not retain raw account emails, accept private study notes, provide public user profiles, or accept new research-file uploads. Supabase retains the authentication account required to sign you in. No online service can guarantee absolute security.

09

Your rights

Where POPIA or another applicable law applies, you may ask whether the Operator holds your personal information; request access, correction or eligible deletion; object to certain processing; withdraw consent where consent is the basis; or lodge a complaint with South Africa’s Information Regulator.

The Site uses automated decisions for admission, classification and note screening. You may revise and resubmit an idea. Requests for any legally required review or redress may be sent to the privacy contact below.

10

Cookies, authentication, analytics and advertising

The Site uses only essential cookies or similar browser storage needed for passwordless authentication, security and recording the accepted Terms version. If you save ideas, the numbers of the ideas you saved are kept in your own browser so you can find them again; they are not sent to the Site or linked to an account, and clearing your browser storage removes them. The anonymous admission totals described above are operational quality metrics, not behavioural analytics: they do not track readers, accounts, journeys or individual submissions. The Site has no advertising cookies, tracking pixels or user-profile database. To understand how the Site is used, it counts visits with Cloudflare Web Analytics, which sets no cookies, does not follow you across other sites and does not build a profile of you. It records the page visited, the referring site, and the general country, browser and device type, and shows the operator only totals. Introducing advertising, or analytics that identify or follow individual visitors, would require a new review, an updated notice and any necessary choices or consent.

11

Age

Public scholarly content may be read without using the AI tools. AI-assisted contribution tools are restricted to people aged 18 or older. Research Idea does not knowingly invite children to submit personal information.

12

Payments and credit records

PayFast processes payer and payment details for credit purchases and voluntary support. Research Idea does not receive or store card or bank credentials. It keeps a transaction reference, payment kind, amount, status and timestamps. A credit purchase is linked only to the protected account identifier and credit ledger needed to supply the purchase, prevent duplicate grants and reconcile payments.

Payment and ledger records may be retained for consumer protection, reconciliation, tax, fraud prevention and other legal obligations. Support payments do not affect public research records or academic treatment.

13

Changes and contact

This Notice may be updated as the service, providers or law changes. Material changes will be clearly dated and may require renewed acceptance where appropriate.

Privacy, access, correction, deletion, objection and automated-decision enquiries may be sent to contactrir@researchidearegistry.org. Complaints may also be submitted to South Africa’s Information Regulator through its official channels.

!

Minimisation reduces risk

Data protection law still applies to authentication, public citations, automated decisions, provider logs and rights requests. The Operator reviews provider arrangements, retention, cross-border safeguards and accountability requirements as the Site changes.