Privacy Policy

Effective July 5, 2026 · Updated November 5, 2026

NewFamiliar Corp, doing business as Jinn (“Jinn,” “we,” “us,” or “our”), operates jinn.works. This Privacy Policy explains what data we collect, how we use it, and your choices.

1. What We Collect

Brand Analyzer. When you submit a URL to our Brand Analyzer, we collect the URL you provide and the publicly available content on that page. We use this content to generate a brand intelligence report.

Email & Waitlist. When you request access or unlock a report, we collect your email address along with:

  • The page you were on when you signed up
  • Referral source (if any)
  • UTM campaign parameters (if present in the URL)
  • Your timezone, browser language, and screen size

Fama accounts & subscriptions. When you create a Fama workspace or subscribe to a paid plan, we collect your account email, the brands and website domains you add, the custom prompts you submit, and your audit history. Payment details (card number, billing address) are collected and stored by Stripe - we never see or store your full card number.

Automatically. Standard server logs (IP address, browser type, pages visited). We also use cookies and similar technologies for site analytics and for measuring our advertising. Where the law requires it, we ask first. Elsewhere they are on by default, and you can turn them off at any time with the Privacy choices link in the footer.

2. AI Agents and Connected Apps

You can connect an AI agent, such as Claude or ChatGPT, to Jinn. The agent can then read your brand’s record and use Jinn’s tools for you. It connects either by signing in with Jinn or with a token you create in your account.

What we collect.

  • When an agent signs in: your Jinn account ID, the brand it reads, which app connected, and when.
  • When you create a token: its name, the brands it can read, and a one-way fingerprint of it. We never store the full token.
  • Each time an agent uses a tool: an activity log entry with the tool’s name, the brand, the time, whether it worked, how long it took, any cost, and a one-way fingerprint of the request. The log does not keep the text of the request.
  • When an agent asks a question about a brand: the question’s text and which parts of the brand’s record answered it.
  • When an agent creates a brand: the brand’s name and website, and text read from its homepage.
  • When an agent suggests a change to a brand’s record: the suggestion, which waits for the brand’s owner to approve it.
  • When an agent asks us to review an image or brief a task: we send the image or the task to an AI model to answer. We do not save the image or the task’s text.
  • When an agent asks for a picture: the description it sent, your Jinn account ID, the brand, and the pictures that come back.

We keep these records in our Supabase database. A picture request is also held in Upstash, a short-term data store, so the agent can check on the picture. We keep it there even if the picture fails.

Why. To run the tools your agent calls, keep each agent to the brands it may read, apply plan limits and usage caps, show you which agents are connected, fix problems, and show a brand’s owner what their record could not answer.

Who receives it.

  • The brand’s owner. A question an agent asks about a brand can be shown, word for word, to that brand’s owner in their Jinn briefing and the briefing email.
  • WorkOS runs the sign-in when an agent signs in with Jinn.
  • Anthropic’s Claude models, through OpenRouter, read the images you send for review, the tasks you send for a brief, a new brand’s homepage text, and picture descriptions.
  • Picture requests run through n8n, our workflow service. Image models at fal and OpenAI make the pictures. Airtable keeps a record of each picture job.
  • Reading a new brand’s website runs as a background job on Inngest, and may use Browserbase, a hosted web browser.
  • PostHog receives a usage event for each tool call: the tool’s name, timing, the brand, and your account ID. It does not receive what the agent sent.
  • Sentry receives an error report when a tool fails: the tool’s name, the brand, and the error. It does not receive what the agent sent.

How long we keep it. Activity log entries are deleted after 90 days. Entries from attempts that never signed in are deleted after 14 days. Saved questions are deleted after 90 days, but a question already shown to a brand’s owner stays in their briefing and in the email they received. Picture requests are deleted from Upstash after 7 days. Sign-in records, tokens, brands an agent created, pictures it made, and suggested changes are not deleted automatically, even after you disconnect an agent or revoke a token. To have them deleted, follow Your Choices below.

How to disconnect. In your Jinn account, open Settings, then API & Agent Access. Press Disconnect all agents to cut off every agent that signed in with Jinn. They stop reading your brand right away, and we ask WorkOS to end their sign-in. To cut off an agent that uses a token, press Revoke next to that token on the same page.

3. How We Use It

  • To generate and deliver your brand intelligence report
  • To notify you when access is available (waitlist)
  • To understand how our site is used and improve the product
  • To communicate with you about Jinn (only if you opted in)

We do not sell your data. We do not run a newsletter or drip campaign unless you explicitly opt in.

4. Where We Store It

Fama account, subscription, and audit data and Jinn waitlist data are stored in Supabase (managed PostgreSQL, US region). Brand analysis records may also be stored in Airtable (cloud-hosted, US-based). These systems support our analysis pipeline, and we use industry-standard security practices to protect your information.

5. Third-Party Services

We use the following services to operate Jinn and Fama. Each has its own privacy policy, and we share only the minimum data each needs to function.

  • Supabase - account authentication and database
  • Stripe - subscription billing and payment processing
  • Resend - transactional and sign-in email
  • Vercel - application hosting
  • Airtable - legacy Brand Analyzer report storage
  • PostHog - product usage metrics and session replay
  • Google Analytics and Google Tag Manager - site analytics and tag management
  • LinkedIn - advertising measurement
  • Meta - advertising measurement
  • Apollo - business enrichment data
  • DataForSEO - search-result and keyword data
  • WorkOS - account sign-in, including for AI agents
  • Inngest - background job processing
  • n8n - workflow automation for making pictures
  • fal - AI image generation
  • Upstash - short-term data store for usage limits and picture requests
  • Sentry - error monitoring

To measure how AI assistants describe your brand, Fama sends brand-related queries to the following answer engines. These queries contain brand and product terms, not your account details.

  • OpenAI - ChatGPT measurement and text embeddings
  • Anthropic - Claude measurement
  • Google - Gemini measurement; Search Console and Analytics when you connect them
  • Perplexity - Perplexity measurement
  • xAI - Grok measurement
  • OpenRouter - model gateway that routes our Anthropic and xAI requests
  • Browserbase - headless-browser measurement of assistants such as Microsoft Copilot

6. Your Choices

You can request deletion of your data at any time by emailing dan@jinn.works. We will delete your information within 30 days.

7. Changes

We may update this policy as Jinn evolves. Material changes will be posted on this page with an updated effective date.

8. Contact

Questions? Email dan@jinn.works.