Privacy

Privacy policy

A plain-English explanation of what MakerThread™ processes, what we store, and how the browser extension handles message data.

Who we are

MakerThread™ provides an AI drafting assistant for maker-led marketplace sellers. The product includes a Chrome extension, this website, account management, billing, settings, and backend API services.

Contact help@makerthread.app for privacy questions, account requests, or deletion requests.

What information we collect

Account information: your name, email address, authentication identifiers, workspace details, and account preferences.

Subscription information: plan, subscription status, Stripe customer and subscription identifiers, trial status, billing-period dates, cancellation state, and payment-status metadata. Card details are handled by Stripe and are not stored by MakerThread™.

Reply settings: tone settings, staff sign-off names, signatures, FAQs, response templates, business description, policy notes, and other support instructions you choose to save locally in the extension and include with a draft request.

Draft request content: the current Etsy, Faire, NOTHS, or Zendesk conversation text, detected buyer name, detected product title where available, typed support notes, typed prompt text, existing draft text when rewriting, and selected reply settings. This is processed to generate the requested draft.

Voice prompt content: if you use the microphone feature, the audio clip is uploaded to the backend and sent to our transcription provider so it can be converted into text.

Usage and diagnostic metadata: timestamps, channel, request status, model, input/output character counts, error codes, and latency.

Support and enquiry correspondence: sender and recipient addresses, subject, email text, attachment names and sizes, our replies, internal support notes, and delivery and ticket status. Our email provider also processes the original email and any attachments.

Optional signup attribution: your answer to ‘How did you find us?’, any short ‘Other’ answer, and limited first-landing source information. Your self-reported answer is kept separately from automatically detected campaign labels and referring website host.

Website activity and discovery sources

On selected public pages, we count page views by UTC calendar day, page path, referring website host, and bounded utm_source, utm_medium and utm_campaign labels. We do not store full referring URLs, their paths or query strings, search terms, IP addresses, device fingerprints or visitor identifiers in these analytics records. We do not treat page-view totals as a count of unique visitors.

First-landing attribution is held only in memory while this website document remains open, including navigation between its pages. It is lost when the document is reloaded or closed. This analytics feature does not create cookies or use local storage. Authentication separately keeps your sign-in session so the account features work.

If you sign up, the limited first-landing attribution and your optional discovery answer are saved with your account. These answers and campaign labels help us understand acquisition sources; they do not affect your account permissions or access.

For website sign-ins and explicit website sign-outs, we record the event type and time, account identifier and a verified authentication-session identifier for deduplication. Each event type is counted at most once per authentication session. Platform-administrator sessions are excluded. These are website account events, not measurements of time on site, active users, browser closures or extension sign-ins.

When your browser sends Do Not Track or Global Privacy Control, this client skips automatic attribution, page-view reporting and website sign-in/out analytics. You can still choose to answer the optional signup question.

Public-page counts are aggregated rather than stored as individual browsing histories. Analytics is best-effort: privacy signals, blocked scripts, network errors and rate limits can mean some page views or account events are not counted. To bound storage, additional source combinations beyond the daily limit are combined into an ‘other’ group. Hosting providers may maintain separate operational security logs under their own policies.

How the browser extension works

The Chrome extension runs on supported message pages, currently Etsy messages, Faire conversations, NOTHS enquiries, and Zendesk tickets. It reads the visible conversation and reply box on the page you are using so it can prepare a draft request.

The extension does not call OpenAI directly. It sends draft and transcription requests to the MakerThread™ backend, where authentication, subscription checks, usage enforcement, validation, and provider calls happen server-side.

The extension stores local session and reply-setting data in Chrome extension storage so the product can remember your sign-in state, active tone, staff sign-off, FAQ settings, and response templates.

The extension does not use Etsy, Faire, NOTHS, or Zendesk APIs and does not scan your full inbox in the background.

How we use information

To create and manage your account.

To authenticate extension and website requests.

To generate, rewrite, insert, or transcribe draft replies at your request.

To apply subscription access, trials, plan limits, and usage allowances.

To provide billing, cancellation, and subscription-management flows.

To store and apply your reply settings, FAQs, response templates, tones, and staff sign-offs.

To monitor reliability, prevent abuse, diagnose failures, and keep the service secure.

To respond to support requests and comply with legal obligations.

AI providers and model training

MakerThread™ uses OpenAI API services to generate draft replies and transcribe voice prompts. We send only the information needed to complete the request, such as the current message thread, your instructions, and relevant reply settings.

We do not opt in to AI provider training using your draft requests. OpenAI states that API data is not used to train or improve OpenAI models by default unless the customer explicitly opts in.

AI providers may process requests and maintain limited abuse-monitoring or security logs according to their own API terms and policies.

What we do not do

We do not sell customer messages.

We do not use your customer messages to train our own AI model.

We do not expose AI provider API keys in the browser extension.

We do not intentionally store full raw marketplace conversations from draft requests by default.

MakerThread™ does not provide access to browse your marketplace or Zendesk inbox.

We do not store full payment card numbers on MakerThread™ systems.

Logging and retention

Draft requests need to be processed in memory and by our AI provider so the draft can be generated. By default, MakerThread™ is designed to keep usage and diagnostic metadata rather than full raw conversation content.

We keep request outcome metadata and usage records for billing, abuse prevention, and service reliability.

The current staff beta records request metadata rather than raw conversation text. If a support issue requires message content, we will ask you to provide the relevant excerpt directly.

Messages and attachments you send directly to our support or general-contact addresses are kept in our email service and private admin ticket history so we can answer and follow up. If you include a marketplace conversation in your email, that excerpt becomes part of the support correspondence. This inbox does not currently have an automatic deletion schedule. Unsent attachment uploads expire after 48 hours and are removed during subsequent inbox maintenance.

We use OpenAI to categorise support and sales emails, summarise conversations and help our operators draft replies. Only a limited portion of the email conversation is provided; attachments, account diagnostics and private notes are not automatically included. Operators review replies before sending. Cached AI draft results are cleared after one day during inbox maintenance, and the separate usage ledger is retained for up to 30 days until the next maintenance run.

Account, subscription, and usage records may be kept while your account is active and for a reasonable period afterwards where needed for accounting, security, legal compliance, dispute handling, and product integrity.

Support and service-provider access

Authorised operators may access account, subscription, usage, and diagnostic records through MakerThread™'s private admin area and our service providers’ administration tools when needed to operate and support the service.

Operators who enable notifications register a device push subscription with our private admin service. Notification text is generic and does not include customer names, email addresses or message content. The device’s browser or operating-system push service delivers these alerts. Operators can disable notifications in admin settings.

Our admin inbox matches a sender's email address to an account to help investigate their enquiry. It shows recorded request failures and account details alongside the conversation. This does not give operators access to browse the sender's marketplace inbox.

Access to those systems is intended for authorised personnel supporting account requests, security, maintenance, abuse investigation, or legal obligations.

Our application usage records contain request metadata rather than full customer conversations.

Security measures

We use server-side API routes for sensitive logic, so AI provider secrets, Stripe secrets, and service-role keys are not included in the Chrome extension.

We validate draft and transcription requests, apply request-size limits, and route AI access through authenticated backend endpoints.

The extension requests access to its configured MakerThread™ backend and supported Etsy, Faire, NOTHS, and Zendesk pages.

We use managed service providers for authentication, hosting, database, billing, email, and AI processing, and we expect those providers to apply their own security controls.

No internet service can be guaranteed completely secure, but we design MakerThread™ to minimise the amount of sensitive message data we store.

Subprocessors

MakerThread™ uses Vercel for hosting, Supabase for database and authentication services, Stripe for billing, Resend for support and enquiry email, and OpenAI for AI generation and transcription.

Subprocessors may process personal data only as needed to provide their services to MakerThread™.

Your choices and rights

You can change your password through the account flow and manage reply preferences in the extension’s Settings screen.

You can cancel or manage your subscription from the billing area where available.

You can ask us to access, correct, export, or delete personal data associated with your account, subject to legal, accounting, security, and fraud-prevention requirements.

You can uninstall the Chrome extension at any time. You can also clear local extension storage through your browser if you want to remove locally stored session/settings data.

International processing

MakerThread™ and its subprocessors may process data in countries outside your own. Where required, we rely on appropriate contractual and organisational safeguards for international transfers.

Children

MakerThread™ is intended for business use by marketplace sellers and is not intended for children.

Changes to this policy

We may update this policy as MakerThread™ develops. If a change materially affects how we process customer data, we will update this page and, where appropriate, notify account holders.