Privacy Policy

Effective August 1, 2026

1. No Accounts, No Personal Data

Berean.ai does not require or offer user accounts. We do not ask for or store any personal information such as your name, email address, phone number, or physical address. You never need to sign in, register, or identify yourself in any way to use this service.

2. How Your Questions Are Handled

When you ask a question on Berean.ai, several things happen to protect your privacy:

• Your original question is never stored or published. Before an answer is generated, an AI classifier reads your question and extracts a short, anonymous topic description (for example, “biblical view of forgiveness”) and a general category (for example, “Christian life”). This topic description contains no personal details — it is a generic summary of the biblical subject matter.
• Only the anonymous topic and the AI-generated answer are saved. The answer is a general biblical/theological response. It does not reference your original wording, personal situation, or any identifying details.
• Your question text is kept only in a private quality log. The question and the answer are retained in an internal log used for quality review, theological accuracy checks, and abuse prevention. This log is not public, is not linked to your name, IP address, device, or session, and is not the source of the public answer pages described in Section 5. It is never sold, shared, or used to build a profile of you.
• Your question text is sent to AI models for inference. To generate the answer, the question is processed by AI language models — either our own self-hosted models running on Berean.ai infrastructure, or, as a failover when our self-hosted models are unavailable, by Groq, a third-party inference provider. See Section 3 for full details on third-party AI processing.

3. Third-Party AI Processors

To answer your question, Berean.ai sends the text of your question to one of the following AI inference services:

Primary: Self-hosted models. Most questions are answered by AI language models we run on our own infrastructure. In this case, no third party receives your question.

Failover: Groq. When our self-hosted models are temporarily unavailable, your question is sent to Groq Inc. (groq.com) for inference. Groq is a third-party AI infrastructure provider.

• What is sent to Groq: only the text of your question. We do not send your name, email, IP address, device identifiers, location, account information, or any other personal data — because we do not collect any of those in the first place.
• What Groq does with it: under our agreement with Groq and Groq’s published data policies, Groq does not train any AI model on your question or on the answer we generate, and does not retain inputs or outputs by default. Groq may temporarily log inputs and outputs (up to 30 days) only when troubleshooting platform reliability issues or investigating suspected abuse. This is equivalent to the protection Berean.ai itself provides. For details, see Groq’s Privacy Policy and Your Data in GroqCloud.

We do not send your question to any other third-party AI service. Berean.ai does not use OpenAI, Anthropic, Google, or any other consumer-facing AI provider for question answering.

4. Mobile App Consent

Users of the Berean.ai iOS app are shown an in-app disclosure screen before any question is sent for AI processing. The screen identifies the AI processors named in Section 3, explains what data is sent, and requires the user to tap “I agree” to proceed. Consent can be revoked at any time from the app’s Settings screen, after which AI features will be disabled until consent is re-granted.

5. Published Answer Pages (/q/ URLs)

Berean.ai publishes answer pages at URLs like /q/biblical-view-of-forgiveness. These pages are designed to help others find answers to common biblical questions through search engines. Here is how this system protects your privacy:

• Topic titles are AI-generated, not your words. The URL slug and page title come from the AI classifier’s anonymous topic extraction — not from your original question. If you ask “my wife and I are struggling to forgive each other after what happened last year,” the resulting page might be titled “What the Bible says about forgiveness in marriage” with no trace of your personal situation.
• Answers contain no personal information. The published answer is a general biblical/theological response about the topic. It is not a reply to you personally — it reads like an article on that subject.
• Individual answer links are random IDs. When you share an answer, the URL includes a short random identifier (e.g., /q/biblical-forgiveness/a3f8b2). This ID is a random string with no connection to you, your device, your IP address, or your session.
• No IP address, device fingerprint, location, or session data is stored with answer pages. There is no way to trace a published answer page back to the person who asked the question.
• Non-canonical answers are automatically deleted. When multiple people ask about the same topic, only the first answer is kept as the permanent page. Additional answers are temporary and are automatically removed after 90 days.

6. Recommend for Publication

After you receive an answer, you may optionally click Recommend to suggest that question and answer for publication on Berean.ai. This is entirely voluntary.

• What happens when you recommend: we flag that Q&A pair for editorial review. Editors may consider it for a public answer page (see Section 5).
• What is retained for review: the question text and AI-generated answer associated with that recommendation, so editors can evaluate quality and appropriateness.
• What is published: if selected for publication, the public page still uses an AI-generated anonymous topic title and a general theological answer — not your personal wording as a byline, and not any identifying information about you.
• No account required. Recommending does not create an account, and we do not ask for your name or contact information to submit a recommendation.
• You can simply not click Recommend. If you do not recommend a Q&A, it is not flagged for this publication-review queue.

7. Follow-up Conversations

After you receive an answer, you may ask follow-up questions in the same conversation. Follow-ups are optional — you can ignore them entirely and simply start a new question.

• The conversation lives in your browser. The running thread of questions and answers is held in your browser tab only. We do not store conversation threads on our servers, and there is no way to resume a conversation later.
• Refreshing or clicking “Ask a New Question” erases the thread. Once the page reloads or you start a new question, the previous conversation is gone from your browser and cannot be recovered by us or by you.
• Recent context is sent with each follow-up. So that the answer makes sense in context, the most recent messages in the conversation are sent along with your follow-up to the same AI processors described in Section 3. The same protections in Section 3 apply.
• Follow-ups are never published. Follow-up questions and answers do not create public answer pages (Section 5) and are not eligible for the Recommend queue (Section 6). Only the first answer in a conversation can ever become a public topic page.
• Follow-ups appear in the private quality log. Like other answers (Section 2), follow-up questions and answers are recorded in our internal quality-review log, flagged as follow-ups. This log is not public and is not linked to your identity.
• Conversations are length-limited. A conversation allows an opening question plus a small number of follow-ups, after which you are asked to start a new question.

8. What We Do Not Collect

To be explicit, Berean.ai does not collect or store:

• Your name, email, or any contact information (unless you voluntarily submit the contact form)
• Your original question text for publication, except when you voluntarily recommend a Q&A for publication (Section 6)
• Your IP address in connection with questions or answers
• Device fingerprints or unique identifiers
• Location data
• Browsing history or tracking cookies (beyond Google Analytics, described below)
• Any information that could identify you as the person who asked a particular question

9. Local Storage

Your browser or mobile app may use local storage to remember preferences (such as conversation state or your AI-processor consent choice). This data is stored locally on your device and is never transmitted to our servers.

10. Analytics

We use Google Analytics to understand how the service is used in aggregate (page visits, browser type, time on site). Google may use cookies and collect anonymized usage data. You can opt out using Google’s browser add-on. Analytics data is not linked to individual questions or answers.

11. API & MCP Usage

When you use the REST API or MCP server, we log the request for rate limiting and abuse prevention. API requests are tracked by IP address for rate limiting only — IP addresses are not stored long-term and are not linked to the content of questions or answers. Questions submitted via the API and MCP server are processed by the same AI processors described in Section 3.

12. Scholar Mode

Scholar Mode queries are logged with the question text for quality improvement purposes. Scholar Mode does not publish answer pages. These logs do not include IP addresses or personal identifiers. Scholar Mode queries are processed by the same AI processors described in Section 3.

13. Other External Services

In addition to the AI processors described in Section 3, Berean.ai may link to or embed content from external services (Bible verse popups via RefTagger, Google Fonts, Google Analytics) which are governed by their own privacy policies.

14. Data Security

All traffic is served over HTTPS. We take reasonable precautions to protect site integrity, but no website is completely secure.

15. Children

Berean.ai is not directed at children under 13. We do not knowingly collect data from children.

16. Changes

We may update this policy from time to time. The effective date at the top of this page will be updated. Continued use of the service constitutes acceptance of any changes.

August 1, 2026 update: Added Section 7 (Follow-up Conversations) describing multi-turn follow-ups on the question page: the thread is held in your browser only, recent context is sent with each follow-up, follow-ups are never published as answer pages, and conversations are length-limited. Corrected Section 2, which previously said question text was discarded and never written to storage; questions and answers are in fact retained in a private, non-public quality-review log that is not linked to your identity. Subsequent sections renumbered.

July 27, 2026 update: Added Section 6 (Recommend for Publication) describing the optional Recommend button on answers, what is retained for editorial review, and how publication still uses anonymized topic titles. Clarified the “What We Do Not Collect” section regarding original question text when a recommendation is submitted.

May 20, 2026 update: Added Section 3 (Third-Party AI Processors) naming Groq as our failover inference provider and identifying the data sent, the recipient, and the contractual protections that apply. Added Section 4 (Mobile App Consent) describing the in-app disclosure and consent flow used by the Berean.ai iOS app. Minor wording updates throughout to reflect that “the site” is now “the service” (web + mobile app).

17. Contact

Questions about this policy? Contact [email protected].