Privacy Policy

RECEIPTS is a U.S.-focused legal-information and research website. This policy describes the data flows in the current free public beta. New public payments, subscriptions, and AI research remain on hold during this beta.

1. Data kept on your device

RECEIPTS uses browser storage for bookmarks, notes, recent searches, saved Receipts, drafts, accessibility/display preferences, return-to-task state, and—after you use account sign-in—the last sign-in email on that device so RECEIPTS can offer a faster returning-user sign-in. This data stays in that browser unless you clear it, export it, or—after sign-in—use account work sync or backup features.

Each submitted question is saved as its own search in this browser, with the answer, sources, and research location when available. Search history stays collapsed under Saved, and previous questions are not shown on Home or shared page headings. Clear screen hides the current question and answer while keeping search history. This controls what appears on screen; it does not encrypt browser storage or delete saved research. Use Clear search history or the device-data controls to remove stored searches. Search history is excluded from automatic account sync unless you select it in a manual sensitive-data backup.

Optional community feedback is sent to the private site inbox when you press Send feedback. You may select the step you were using and whether the issue prevented you from continuing. An email for a reply is optional. Your current question and research location are excluded unless you select Include my current question. The form does not send your search history, saved answers, or account session. Feedback is kept only as long as reasonably necessary to review and respond to it. The owner can review these reports privately and copy a summary of reported steps, impacts, review statuses, submission dates, and feedback references; that summary omits report text, questions, documents, and reply emails.

The source-catalog finder searches existing public source descriptions in your browser. Its search text is not placed in a URL, saved by the finder, or sent to Google or the RECEIPTS server. Question interpretation for source routing also runs locally. Other research, saved-work, account-sync and feedback features follow their own rules below. Opening an external source visits that publisher’s website.

1A. Local document scanning

When you take a photo, choose an image, or select a PDF in the basic document reader, RECEIPTS processes that selected file in your browser. Image reading downloads an OCR/text-recognition engine and PDF text extraction downloads a PDF reader. The selected file is not uploaded to the RECEIPTS server for this basic scan. Extracted text can contain mistakes, so compare it with the original before relying on it. If you later send extracted text into another RECEIPTS feature, that feature follows the data rules described for that feature in this policy.

2. Optional accounts and authentication

If you choose Continue with Google, Google and Supabase process the sign-in and RECEIPTS receives the basic account identity needed to authenticate you, such as your verified email and account identifier. RECEIPTS does not request access to read your Gmail inbox, contacts, or messages. The secure email-link method remains available as a backup. Public research does not require an account, and RECEIPTS does not ask you to create a site password.

3. Account work sync and optional backup

Signed-in users can keep a privacy-minimized copy of bookmarks, personal notes attached to bookmarks, saved Receipts, and reading/display preferences in a Supabase row protected by Row Level Security so that saved work can follow the same account to another device. Search history and drafts are excluded from automatic sync by default. They are included only when you affirmatively select the manual sensitive-data backup option.

Account controls let you export or delete the cloud backup. Deleting the RECEIPTS account is designed to remove the Supabase authentication account and cascade-delete RECEIPTS account rows. Historical payment records held by Stripe may remain where required for payment, accounting, fraud-prevention, dispute, or legal purposes.

4. Payments and subscriptions

Stripe processes the optional one-time support payment and any premium subscription a customer deliberately activates through public checkout. RECEIPTS does not receive or store a full payment-card number. RECEIPTS may keep the minimum Stripe customer/subscription identifiers and status needed to confirm access, prevent duplicate billing, and open Stripe's subscription-management tools.

5. Hosting, security, and technical data

Cloudflare hosts and secures the site and may process ordinary request, network, device, performance, and security metadata needed to deliver the service. RECEIPTS does not intentionally embed third-party advertising trackers in the current build.

Cloudflare Turnstile may receive the visitor's network address and security/browser signals to perform a bot check for anonymous community feedback and, when enabled, paid AI research. The feedback security check verifies the submission; it does not add the visitor's network address to RECEIPTS' aggregate flow counters. RECEIPTS' own daily usage counter is designed to store a keyed pseudonymous account identifier and numeric request counts, not the raw network address or legal-question text.

If aggregate flow measurement is enabled, RECEIPTS keeps daily counts of section opens, research actions and outcomes, source-link clicks, completed device saves, checkout opens, browser errors, page-asset loading errors, saved-research read/write problems, and read-aloud failures or unavailable device support. Research outcomes distinguish explanations, source-only packets, unresolved results, and fixed categories for holds, access requirements, limits, security checks, setup problems, timeouts, and unavailable service. Only a fixed action name, a limited campaign category, a UTC date, and an owner-test/public category are recorded. Shared links may add a fixed Reddit, Facebook, or Instagram label to a small set of action counters. The full tracking query and arbitrary labels are not recorded. These labels describe the shared link, not a verified visitor origin or an individual; older unlabeled links cannot be assigned a platform later. These counters do not include legal questions, document text, notes, source URLs, error messages, account identifiers, exact event times, or visitor/session identifiers. They cannot identify visitors or reliably separate signed-out owner checks from other public activity. This feature creates no analytics cookies or browser storage, respects Global Privacy Control and Do Not Track signals, and does not replay screens or identify people across visits. Reports cover the last 90 days; older counter rows are removed on the next successful collection after their reporting window ends. Counts are best effort: blocking, offline requests, older app versions, and browser crashes can prevent recording. Opening an external source link does not establish that the destination loaded.

Error categories may also be grouped by the broad site section where they were observed, such as Research, Saved, or Learning. This does not include a research topic, individual question, account identity, or a visitor's sequence across visits.

The same optional counters include repeated taps on selected primary controls (three within 1.5 seconds), viewport depth reaching 25%, 50%, or 90% in Home, Research, Learning, or Documents, question editing started or left before a recorded submission, and optional Yes/No helpfulness choices grouped by broad result type. Only fixed categories reach the counter service. Input detection checks whether an edit occurred in a known question field; it does not read or send typed characters, keystrokes, field contents, or field lengths. Password and account fields are excluded. Tap timing and element references stay temporarily in memory and are not sent. Pointer coordinates and mouse movements are not collected. These signals do not prove confusion, reading, abandonment, or satisfaction: automatic scrolling can affect depth, drafts may be resumed, and helpfulness responses are self-selected. Additional signals are limited to 30 per page load within the existing collection limits. They add no visitor profiles, recordings, cookies, or browser storage.

During beta, the server also copies the last seven UTC days of these existing daily counts into one private summary in Supabase for authorized operator review. Public activity and authenticated owner tests remain separate. Each successful copy replaces the previous summary; it does not create a new event history. The copy includes a report-generation timestamp, not individual event times, and expires after seven days. An expired copy is removed on the next successful copy. Delivery is best effort, so an older or unavailable summary does not mean there were no visitors or problems. Ordinary visitor and customer accounts cannot read this private summary.

6. Live AI research

Free rule cards and local help tools can operate without sending a legal question to an AI provider. Only when a signed-in authorized premium user deliberately starts live AI research and the feature is enabled is the question and selected research context sent through the RECEIPTS server to the OpenAI API. The request is made with model storage disabled by the application. OpenAI states that API inputs and outputs are not used to train its models by default; provider-side retention, abuse monitoring, security, and legal obligations are governed by OpenAI's current business/API terms and policies.

OpenAI enterprise/API privacy information ↗

7. External official-source links and software delivery

When you open an external government, court, agency, or other source website, that destination receives the request and applies its own privacy practices. The current site loads the Supabase browser library from jsDelivr and may load Cloudflare Turnstile only when that security feature is needed; those providers can receive ordinary connection metadata.

8. Legal, privacy, copyright, and accessibility requests

If you submit the RECEIPTS privacy/legal request form, RECEIPTS processes the contact information and message you provide to review and respond to that request. Do not put full financial account numbers, passwords, Social Security numbers, or unrelated case documents in the form. Request records should be kept only as long as reasonably needed for the request, compliance, disputes, security, or legal obligations.

9. How RECEIPTS uses information

• Provide requested site features and authentication.
• Save data only when you choose local or cloud storage features.
• Process payments and subscription status.
• Protect the site against abuse and control paid-AI costs.
• Respond to privacy, copyright, accessibility, and payment issues.
• Comply with legal obligations and protect rights, safety, and security.

10. Sale, sharing, targeted advertising, and Global Privacy Control

RECEIPTS does not sell personal information or use it for targeted advertising in the current build. It also does not intentionally share personal information for cross-context behavioral advertising. The optional aggregate flow counter honors Global Privacy Control and Do Not Track signals as described above. Any future sale, sharing for advertising, or targeted advertising would require applicable notices and controls before it begins.

11. Retention

12. Your choices and privacy rights

Depending on where you live, privacy law may give you rights to request access, correction, deletion, or a portable copy of certain personal information, and rights relating to sale, sharing, targeted advertising, or sensitive data. RECEIPTS provides self-service controls for local data and signed-in cloud backups and accepts privacy requests through the privacy/legal request page. RECEIPTS may need to verify a request before acting on account data.

Make a privacy/legal request

13. Children

RECEIPTS is a general-audience U.S. legal-information service and is not directed to children under 13. Children under 13 should use only the non-account public information with a parent or guardian and should not create an account, make a payment, or submit personal information. If RECEIPTS learns that personal information was collected online from a child under 13 in a manner covered by COPPA, it should be deleted or handled in accordance with applicable parental-consent requirements.

FTC COPPA information ↗

14. Security and breach response

RECEIPTS uses HTTPS, account authentication, Row Level Security for the backup table, server-side secret handling, input limits, and security headers. No internet service can promise perfect security. If a reportable data incident occurs, RECEIPTS intends to investigate, contain it, and provide notices required by applicable law.

15. U.S. focus and processing

RECEIPTS is designed around U.S. public-law research and is not marketed as an international legal service. Service providers may process data in the United States and other locations described in their terms. Access from another country may create additional legal obligations that should be assessed before actively marketing there.

16. Changes

This policy should be updated before materially changing data collection, adding advertising technology, changing document processing from local/browser-only scanning to server upload or storage, or introducing a new provider or new use of personal information.