1. Operator and contact
PathwayFig is operated by an individual developer based in China. This policy covers personal information processed by PathwayFig when you use the website, accounts, AI pathway analysis, optional AI image generation and support/report channels. For privacy questions and requests, contact [email protected].
This policy describes the current beta, including one-time credit purchases processed through Waffo Pancake. External sites and service providers have their own privacy terms; a link here does not make their processing identical to ours.
2. Information we process
| Information | Use and location |
|---|---|
| Vector edits, drafts and exports | Vector editing uses browser page state. Unsaved changes are copied to session storage in the current tab so a reload can offer to restore them; this copy is removed when the tab closes, when you discard it or when the account in that tab changes. Edits are not uploaded unless you choose Save to account. You can also save a JSON draft to your device and reopen that file; opening a file does not upload it. Export SVG or PNG for use elsewhere. |
| Diagrams saved to your account | When you are signed in and choose Save to account, the diagram title, node names, types, positions, colors and descriptions, relationships and labels are stored with your account ID, a revision number and created/updated timestamps in our database. These saved diagrams are visible only to your signed-in account. Each account can keep up to 50, each limited to the same size as an editable file. Deleting a diagram in My diagrams removes it from the database. Saving does not send the diagram to an AI provider. |
| Browser preferences and pending tasks | A theme preference can be saved in local storage. Pending AI request IDs can be saved in session storage to resume status checks. Legacy local records may remain on your device after password fields are removed. |
| Account information | Email address, display name, salted password hash, session records, policy acknowledgments and access restrictions are stored on the server when accounts are enabled. We do not store plaintext passwords. If you choose Google sign-in when available, we verify a Google identity token and store your Google account identifier, email and display name; we do not receive your Google password or store the identity token. |
| Account recovery | When email recovery is enabled, we send your email address and a one-time reset link to Resend to deliver an account email. Our database stores a hash of the reset token, the related account, expiry and use timestamps; it does not store the raw token. The link expires after 30 minutes and a successful reset invalidates other reset links and existing sessions. Successful recovery records that mailbox ownership was verified. |
| AI graph analysis | Your description passes through our server to APIMart for structured extraction and content screening. We store an account-linked request hash, status, timestamps, usage reservations and the extracted graph, including labels and review notes. Repeating the same description on the same account in the same UTC day reuses the original result. This cache is not an autosave of later edits or a permanent graph library. The raw full description is not stored in this request record, but the extracted graph can contain information from it. |
| AI image inputs and task records | Your prompt and settings pass through our server to the disclosed AI service. Task IDs, request hashes, status, timestamps, resulting image URLs and credit adjustments are stored for delivery, balance integrity and support. Approved generated images are copied to private Cloudflare R2 storage when available, so you can reopen and download them from your account library. Unarchived provider links can expire; download a copy when the library shows storage pending. The library does not currently store the full original prompt. |
| Checkout and payment records | When you request checkout, your account email, selected product and order reference are sent to Waffo Pancake. We store order status, amounts, currency, credits, provider references and verified payment/refund event metadata for delivery and reconciliation. Payment details are entered with the payment provider, not in PathwayFig prompts or reports. |
| Safety and support information | Report category, description, optional contact email, optional task reference, submitting account ID when available, timestamps and review decisions can be stored. Information you email is also processed to address your request. |
| Technical and security information | Hosting infrastructure processes IP address and request metadata. The application uses derived identifiers and counts for rate limits and abuse prevention. During a limited free pilot, it retains an account-linked trial claim, an account-scoped grant reference, grant timestamp and generation-policy acceptance, and attempt/budget records to prevent repeat grants and reconcile provider costs. Earlier trial records may retain a salted network identifier, but sharing a network no longer prevents another account from receiving its trial. Normal error responses do not expose prompts, passwords, API keys or database records. |
Do not send patient records, special-category personal data, confidential research, passwords or payment-card information in prompts or reports.
3. Why information is used
We use information to provide requested accounts, graph analysis and image generation, preserve credit balances and request consistency, operate and secure the service, prevent abuse, investigate content reports, respond to support requests and meet applicable legal obligations.
Where a law requires a legal basis, the relevant basis may be providing the service you request, legitimate interests in operating and protecting the service (subject to your rights), compliance with law, or consent where required. We do not treat a privacy-policy acknowledgment as blanket consent to unrelated processing.
We also use Google Analytics 4 to understand website visits and usage. Google may process browser and device information, approximate location and cookie identifiers. We disable Google signals and advertising personalization, omit URL query strings and fragments from the configured page URL, and do not load the analytics tag when your browser sends Do Not Track = 1. We do not send prompts, account identifiers or form contents to Google Analytics. See Google’s privacy policy for its processing practices.
We collect limited first-party event counts for actions such as viewing a page, opening the editor, exporting, signing up and starting checkout. These event records contain an event type, time, a random event identifier for duplicate prevention and limited settings such as format or resolution. They do not include your account identifier, email, prompt, image URL, IP address or a persistent tracking identifier. The application respects a browser Do Not Track setting of 1 by not sending these events. Event records are scheduled for removal after approximately 30 days and are used for directional usage counts, not unique-person tracking. Payment confirmation and revenue are checked separately against server order records. We do not send these events to advertising services, sell personal information or use it for targeted advertising.
4. Service providers and disclosure
- Cloudflare: hosting, request handling and, where configured, the D1 database that stores account, saved-diagram, graph-analysis, image-task, payment and report records and private R2 storage for generated images. Its infrastructure processes relevant connection metadata.
- APIMart and upstream AI providers: your description and extraction instructions for editable-graph analysis through
gpt-4.1-mini-2025-04-14; image-generation prompts, settings and task processing throughgpt-image-2.5-ext, versionflare. The model disclosure explains the route and limitations. - Content safety: prompt content, returned graph text and generated image references are processed for safety decisions before generation or display. The configured content-safety provider is APIMart (omni-moderation-latest).
- Image hosting: archived images are served through authenticated PathwayFig endpoints; viewing an unarchived AI result may connect your browser to the provider's image URL, revealing ordinary connection metadata to that host.
- Google sign-in (optional, when enabled): the account dialog loads Google Identity Services to offer authentication. Google receives ordinary connection metadata and handles account selection under its own privacy terms. PathwayFig requests identity information only, not access to Gmail or Drive.
- Waffo Pancake: account email, selected product and order references are sent when you request production checkout. Waffo processes any payment information you provide on its checkout, and sends payment/refund confirmation data to PathwayFig where applicable.
- Resend (when email recovery is enabled): your account email and the reset email content, including its short-lived link, are processed for delivery. This feature stays unavailable until the sending service is configured. Resend processing is subject to its own applicable terms.
- Support communications: email services process correspondence you send to the support address.
We may disclose relevant information when necessary to comply with law, respond to a valid legal request, protect users from harm, investigate abuse or resolve a rights claim. Report information is not a public feed. We limit disclosure to what is appropriate for the purpose.
Waffo Pancake processes checkout requests and sends payment and refund confirmations to PathwayFig. A checkout error does not establish whether a charge occurred; check the order status before starting another purchase. PathwayFig does not receive your full payment-card details; consult the payment provider’s privacy information before entering them.
5. International processing and third-party terms
The operator is based in China, and internet infrastructure and AI providers may process information in other countries. We do not currently promise country-specific storage or zero data retention by an upstream provider. Third-party provider handling follows the arrangements and terms applicable to their services.
We do not train our own AI model on user prompts or outputs in this version. That statement is not a guarantee about every third party's training or retention practices. Do not submit information that your institution prohibits you from sharing with external AI services.
6. Retention and security
Browser page state and the tab-scoped recovery copy last for the current tab; locally saved preferences and legacy records remain until you or the browser remove them. Diagrams saved to your account remain until you delete them in My diagrams or ask support to delete your account data; deleted diagrams may persist in database backups for the backup retention period. Session cookies expire after seven days unless you sign out earlier. Authentication, credit, graph-analysis, image-task, payment and report records are retained for service operation, troubleshooting, unresolved disputes and applicable legal obligations.
Same-day reuse of graph analysis is a request-deduplication window, not a promise that the stored graph is deleted at midnight. This beta does not run a general automated deletion schedule for all database records. Ask support to delete data you no longer want retained. We will review the request, verify identity when needed, remove or anonymize information that is no longer required, and explain any applicable reason for retaining some records. Third-party copies and backup retention may follow separate schedules.
We use protections including server-side secrets, password hashing, restricted session cookies and request limits. No internet service or storage system is completely secure. Tell us promptly if you suspect a security issue and avoid sending secrets in a report.
7. Your choices and privacy requests
Depending on the law that applies to you, you may have rights to access, correct, delete or receive a copy of your data, object to or restrict certain processing, withdraw consent where processing relies on it, and complain to an appropriate supervisory authority. These rights can have legal limits and exceptions.
Email [email protected] with the type of request and the account email or relevant reference. Do not include your password. We may need proportionate information to confirm ownership, and we will respond within any period required by applicable law. Account deletion remains a support request. The sign-in dialog shows automated email recovery only when its delivery service is enabled; otherwise, contact support for recovery.
You can use the free editor without an account, decline to submit AI prompts, delete diagrams you saved in My diagrams, sign out, and remove local site storage using browser settings. Clearing local data may remove preferences and interrupted-task recovery information.
8. Children and policy changes
The service is intended for adults aged 18 or older. We do not knowingly solicit children's personal information. If you believe a child has provided personal information, contact support so that we can investigate and take appropriate action.
The effective date appears at the top. We will update this policy when processing changes and provide additional notice or seek consent when required. See Terms of Service, Acceptable Use and Contact.