WARM · PLAIN ENGLISH · LAST UPDATED: OCTOBER 22, 2026
What I do with your personal information
I’m Benjamin Tan, and I run ben-tan.com, my portfolio. This notice covers the whole portfolio.
That includes the pages you browse, artwork, guides, tools, and the Bending Machine. Shared practices come first, then the details for each feature.
What happens across the portfolio
First, the practices shared across the portfolio. Then, the tools and features you choose to use. These illustrations are interactive demonstrations.
Layer 1 · Across the portfolio
01 · BROWSE
Browsing the portfolio
Opening a page sends Cloudflare your IP address, the page address, and other technical request details. It uses them to deliver content and help protect the portfolio.
I use Plausible across the portfolio to measure visits, time spent, and scrolling. It doesn’t receive tool inputs, file contents, or results.
If you email me, I receive your email address, message, and attachments and use them to reply and follow up. Email services deliver and store our correspondence.
I normally delete routine emails 12 months after our last meaningful email exchange. I keep messages longer for ongoing work or required records, and review those exceptions once a year.
When available, Test Data Generator and PII De-identifier process your selected inputs in your browser. The tools don’t send inputs, file contents, or results to me or include them in analytics. Optional saved settings are explained in the tool section below.
Downloading hosted artwork or my resume asks Cloudflare to deliver the file. I don’t track download-link clicks.
The Bending Machine keeps a small record in your browser. This is a demonstration; it doesn’t delete your actual record. Open the Bending Machine to use the real control.
Enter 53 on the Bending Machine’s keypad, then press the # button labeled ENTER. One administrative beverage, please.
STEP 02 OF 04
Collect your drink
Collect the can from the tray.
STEP 03 OF 04
Scan the barcode
Scan the barcode. The Bending Machine prints a deletion request; nothing is deleted yet.
STEP 04 OF 04
Press DELETE
On the receipt displayed by the Bending Machine, DELETE removes the Bending Machine’s browser record from this browser. Loading or restarting the Bending Machine creates a new one. Deleting the browser record doesn’t delete records held by Cloudflare, Plausible, or SomaFM, or emails.
Demonstration, step 1 of 4: Select item #53. Enter 53 on the Bending Machine’s keypad, then press the # button labeled ENTER. One administrative beverage, please.
The full notice
1. The whole portfolio
This notice covers ben-tan.com: its portfolio pages, artwork, guides, tools, interactive experiences, and privacy pages. The practices below are grouped by activity, with feature-specific details available where needed.
If collection, use, storage, or disclosure changes, I’ll update this notice and its date and explain relevant changes alongside the feature.
2. Hosting and analytics
Hosting and security
Cloudflare hosts, delivers, and protects the portfolio. Page and hosted-file requests send it your IP address, requested address, and other technical request details so it can deliver content and help prevent abuse.
I use Plausible across the portfolio, including pages, tools, guides, and interactive experiences. It reports page views, time spent, scroll depth, and the referring site. Download clicks, external-link clicks, and form submissions aren’t tracked.
Plausible uses your IP address and browser details to estimate your location and count visitors within a day. Reports can include approximate location, browser, and device type. It doesn’t keep your IP address or the full browser description sent with the request. It keeps records of events and visits to calculate the overall reports I see. The daily visitor count is separate from how long those records are kept.
Anything after ? or # in a page address is left out of the reports. A referral is reported as just the site address. Tool inputs, file contents, and results are left out. Plausible doesn’t use cookies or save a visitor ID in your browser.
I don’t sell your personal information or share it to target ads across different sites. You don’t need to press anything to stop those activities.
When you email me, I receive your email address, message, and attachments. Email services deliver and store our correspondence. I use what you send to reply and follow up.
I normally delete routine emails 12 months after our last meaningful email exchange. I keep messages longer when needed for ongoing work or required records, and review those exceptions once a year.
The 12-month email period doesn’t apply to hosting, analytics, or radio records. Those records follow the relevant service arrangements and settings; the linked provider policies describe their practices. Plausible lets me delete the site’s analytics. An email address doesn’t let me reliably identify a particular visitor’s analytics records.
You can clear data saved by this site through your browser settings. Browser privacy tools may block analytics requests. Clearing browser data doesn’t erase email, downloaded files, or records already held by Cloudflare, Plausible, or SomaFM.
For questions or requests about information I hold or control, including access, correction, or deletion, email hello@ben-tan.com. Tell me which feature and information you mean. I’ll explain what I can identify and what action I can take. For information kept only in your browser, use the relevant feature controls or browser settings.
Links to LinkedIn, project sites, and other destinations lead to separate services with their own privacy notices.
4. Processing files and choices
Test Data Generator creates sample data and reads selected setup files in your browser. PII De-identifier processes selected files in browser memory. These tools don’t send inputs, file contents, or results to me or a processing server. Your browser creates output and setup downloads when requested. Loading the tools still makes the hosting and analytics requests described above.
Guide searches and choices stay in this tab’s memory. They aren’t uploaded or saved between visits. Reloading restarts the path; Copy puts the result on your clipboard when you ask.
Saved records and downloaded files have separate controls below.
5. Saving work on your device
Some features save records in this browser between visits. Reloading doesn’t clear them. That’s different from temporary tab memory and downloaded files.
Study progress, saved notes and reflections, and separate drafts.
Saved rules, machine history, and CIPT records don’t expire on a timer. Each feature’s controls affect its own records; clearing this site’s browser data is broader. Your browser can also lose or remove stored data. Downloaded files and copies in another browser or at another address are separate.
CIPT records aren’t encrypted or protected by a Studio sign-in. Other scripts on the same portfolio origin can access them. Studio doesn’t send these records to me or automatically sync them across browsers or devices.
CIPT backup files are readable and unencrypted. The progress JSON leaves out separate drafts; draft recovery text is a separate file.
Import replaces progress instead of combining histories. If saving fails, your original saved progress and drafts stay intact. Optional tab-only recovery doesn’t save progress or draft changes; reload brings back the original saved copy.
Turning on the optional radio connects your browser directly to SomaFM. It receives your IP address and the request details needed to stream music. Turning it off stops the stream. It doesn’t delete information SomaFM already received.
Cloudflare delivers hosted files such as artwork and my resume. Tool outputs, setup files, and the machine’s record export are created locally when requested. Download clicks aren’t tracked in analytics. Clearing browser storage doesn’t remove files already downloaded; manage those wherever you saved them.
You can copy or download a CIPT voice-practice brief to share elsewhere. Studio doesn’t use your microphone, record audio, call an AI or speech service, or upload transcripts.
Pages can use temporary browser state to keep navigation working. EMA search text can appear in the page address and history, a link you copy, or a request to the host. Plausible leaves it out of analytics.
Email services deliver and store correspondence. I receive email addresses, messages, and attachments to respond and follow up.
I normally delete routine correspondence 12 months after the last substantive exchange. Ongoing work or required records may require longer retention; I review those exceptions annually.
Working browser-based tools process selected inputs locally. Tool inputs, file contents, and generated results are not transmitted to me or included in analytics. Optional saved settings are described in § 4.
Hosted file requests reach Cloudflare. Download-link clicks are not tracked in analytics. Locally generated tool outputs, setup files, and machine-record exports remain distinct from hosted downloads.
Enabling the optional Bending Machine radio connects the browser to SomaFM, which receives the IP address and stream-request details. Disabling it stops streaming but does not erase information already received.
Enter item number 53 on the Bending Machine’s keypad, then press the # button labeled ENTER.
STEP 02 OF 04
Collect the can
Collect the can from the dispensing tray.
STEP 03 OF 04
Scan the barcode
Scan the barcode to print a deletion request. This step does not delete the record.
STEP 04 OF 04
Press DELETE
Pressing DELETE on the receipt displayed by the Bending Machine erases the Bending Machine’s browser record in this browser. Loading or restarting the Bending Machine creates a new record. Records held by Cloudflare, Plausible, or SomaFM, and email are unaffected.
Demonstration, step 1 of 4: Enter code 53. Enter item number 53 on the Bending Machine’s keypad, then press the # button labeled ENTER.
The full notice
1. Portfolio scope
This notice covers ben-tan.com: its portfolio pages, artwork, guides, tools, interactive experiences, and privacy pages. The practices below are grouped by activity, with feature-specific details available where needed.
If collection, use, storage, or disclosure changes, I will update this notice and its date and explain relevant changes alongside the feature.
2. Hosting, security, and analytics
Hosting and security
Cloudflare hosts, delivers, and protects the portfolio. Page and hosted-file requests send it your IP address, requested address, and other technical request details so it can deliver content and help prevent abuse.
I use Plausible across the portfolio, including pages, tools, guides, and interactive experiences. It reports page views, time spent, scroll depth, and the referring site. Download clicks, external-link clicks, and form submissions are not tracked.
Plausible uses IP addresses and User-Agent information to derive approximate location, browser/device information, and a daily visitor identifier. It does not store raw IP addresses or User-Agent strings. It stores event and session records to calculate the aggregate statistics shown in its dashboard. A daily identifier is not a promise that analytics records are deleted each day.
Reported page addresses exclude query strings and fragments. Referral addresses are reduced to their origin. Tool inputs, file contents, and generated results are excluded. Plausible does not use cookies or browser storage to identify visitors.
I do not sell personal information or share it for cross-site targeted advertising. No control needs to be activated to prevent those activities.
When you email me, I receive your email address, message, and attachments. Email services deliver and store our correspondence. I use what you send to reply and follow up.
I normally delete routine correspondence 12 months after our last substantive exchange. I keep messages longer when needed for ongoing work or required records, and review those exceptions annually.
The 12-month email period does not apply to hosting, analytics, or radio records. Those records follow the relevant service arrangements and settings; the linked provider policies describe their practices. Plausible lets me delete the site’s analytics. An email address does not let me reliably identify a particular visitor’s analytics records.
Your choices and requests
You can clear data saved by this site through your browser settings. Browser privacy tools may block analytics requests. Clearing browser data does not erase email, downloaded files, or records already held by Cloudflare, Plausible, or SomaFM.
For questions or requests about information I hold or control, including access, correction, or deletion, email hello@ben-tan.com. Tell me which feature and information you mean. I will explain what I can identify and what action I can take. For information kept only in your browser, use the relevant feature controls or browser settings.
Links to LinkedIn, project sites, and other destinations lead to separate services with their own privacy notices.
4. Processing files and choices
Test Data Generator creates sample data and reads selected setup files in your browser. PII De-identifier processes selected files in browser memory. These tools do not send inputs, file contents, or results to me or a processing server. Your browser creates output and setup downloads when requested. Loading the tools still makes the hosting and analytics requests described above.
Guide searches and choices remain in temporary tab memory, without uploads or persistent storage. Reloading restarts the path. Copy places the result on the clipboard only when requested.
Study progress, saved notes and reflections, and separate drafts.
Saved rules, machine history, and CIPT records have no automatic expiration. Feature controls affect their own records; clearing this site’s browser data is broader. Browser storage may also be lost or removed. Downloaded files and copies at another browser or origin remain separate.
CIPT records are unencrypted, without a Studio sign-in. Scripts on the same portfolio origin can access them. Studio does not upload these records to me or automatically sync them across browsers or devices.
CIPT backup files are readable and unencrypted. The progress JSON excludes separate drafts; draft recovery text is a separate file.
Imports replace progress rather than merge histories. If saving fails, original saved progress and drafts remain intact. Optional tab-only recovery does not save progress or draft changes; reload restores the original saved copy.
Turning on the optional radio connects this browser directly to SomaFM. SomaFM receives the IP address and stream-request details. Turning it off stops the stream; it does not erase information already received.
Cloudflare delivers hosted files such as artwork and my resume. Tool outputs, setup files, and the machine’s record export are created locally when requested. Download clicks are not tracked in analytics. Clearing browser storage does not remove files already downloaded; manage those wherever you saved them.
CIPT voice-practice briefs are copied or downloaded only on request, for sharing outside Studio. Studio does not access the microphone, record audio, call an AI or speech service, or upload transcripts.
Page navigation can use temporary browser state. EMA search text can appear in the address and browser history, a copied link, or a request to the host. Plausible excludes it from analytics.
I’m Ben—Benjamin Tan if the paperwork insists. I run the portfolio at ben-tan.com.
This covers the whole portfolio: pages, artwork, guides, tools, and the Bending Machine. Shared practices first; individual features next. Yup, three privacy notices for a single website.
Why? Because somewhere along the way, privacy notices that were meant to provide transparency to users suddenly became legalese walls of text to appease auditors. So, here's my solution. I'll have my cake, I'll eat it, and I'll have leftovers.
Paging self-restraint. Is self-restraint in the room?
A website. Somehow, all this paperwork.
Here’s what happens across the portfolio, followed by the individual features. The illustrations are demonstrations. I’ve made the facts dress for the occasion.
Layer 1 · Across the portfolio
01 · BROWSE
Proof that someone showed up
Cloudflare receives your IP address, requested address, and technical request details to deliver content and help protect the portfolio. It’s doing the actual work while I explain one website in three different voices.
I use Plausible across the portfolio to measure visits, time spent, and scrolling. Tool inputs, file contents, and results stay out. A page view doesn’t tell me whether you liked the page. A cruel limitation, but I shall endure.
Your email address, message, and attachments reach me through email services. I use them to reply and follow up. Yes, including the attachment. You put it there.
I normally delete routine emails 12 months after our last meaningful email exchange, unless ongoing work or required records need them longer. I review those exceptions once a year. The inbox isn’t a mausoleum with Wi-Fi.
Routine email · Normally 12 months after our last meaningful email exchange
When available, the tools process your selected inputs in your browser. They don’t send the inputs, file contents, or results to me or include them in analytics. Saved settings get their own explanation below.
Hosted artwork and resume downloads still ask Cloudflare to deliver a file. Download-link clicks aren’t tracked.
The Bending Machine keeps a small record in your browser. You can delete it via the Bending Machine or through your browser settings.
This walkthrough is a demonstration. This walkthrough doesn’t delete the Bending Machine’s actual browser record. To delete that record, open the Bending Machine and press DELETE on the receipt displayed by the Bending Machine. The animated stamp has a robe and a tiny gavel. The animated stamp still has no powers.
Enter 53 on the Bending Machine’s keypad, then press the # button labeled ENTER. Yes, deleting four fields starts with ordering a drink.
STEP 02 OF 04
Collect your can
Collect the can from the tray. Four fields, and I’ve given them a beverage service.
STEP 03 OF 04
Scan the barcode
Drag the can to the scanner, or click it and choose “Scan barcode”. That prints the request. Scanning alone doesn’t delete anything; we haven’t reached the actual delete button yet.
STEP 04 OF 04
Press DELETE
In the actual Bending Machine, press DELETE on the on-screen receipt to remove its browser record from this browser. Loading or restarting the machine afterward creates a new record. This doesn’t delete analytics, hosting logs, emails, or information held by other services. I put a delete button behind a keypad, a drink, and a barcode scanner. Self-restraint remains unavailable.
Demo complete. This walkthrough has not deleted your actual browser record.
Demonstration, step 1 of 4: Select item #53. Enter 53 on the Bending Machine’s keypad, then press the # button labeled ENTER. Yes, deleting four fields starts with ordering a drink.
Loading or restarting the Bending Machine creates a fresh record. These controls don’t erase analytics, hosting logs, emails, or information held by other services. Clearing your browser is housekeeping, not witness protection.
The full notice
1. An entire portfolio. Paperwork included.
This notice covers ben-tan.com: its portfolio pages, artwork, guides, tools, interactive experiences, and privacy pages. The practices below are grouped by activity, with feature-specific details available where needed.
If collection, use, storage, or disclosure changes, I’ll update this notice and its date and explain relevant changes alongside the feature. I build one new button, and suddenly I’m back here documenting its entire lifecycle. So, please enjoy the button. At least one of us should.
2. Who gets what
Hosting and security
Cloudflare hosts, delivers, and protects the portfolio. Page and hosted-file requests send it your IP address, requested address, and other technical request details so it can deliver content and help prevent abuse.
I use Plausible across the portfolio, including pages, tools, guides, and interactive experiences. It reports page views, time spent, scroll depth, and the referring site. Download clicks, external-link clicks, and form submissions aren’t tracked.
Plausible uses your IP address and browser details to estimate your location and count visitors within a day. Reports can include approximate location, browser, and device type. It doesn’t keep your IP address or the full browser description sent with the request. It keeps records of events and visits to calculate the overall reports I see. The daily visitor count is separate from how long those records are kept.
Anything after ? or # in a page address is left out of the reports. A referral is reported as just the site address. Tool inputs, file contents, and results are left out. Plausible doesn’t use cookies or save a visitor ID in your browser.
I’m trying to see whether anyone uses the portfolio. A page view is evidence that someone arrived. If I start treating it as a standing ovation, please take the dashboard away.
3. The inbox and other jurisdictions
I don’t sell your personal information or share it to target ads across different sites. No ceremonial button press is required to prevent those activities.
When you email me, I receive your email address, message, and attachments. Email services deliver and store our correspondence. I use what you send to reply and follow up.
I normally delete routine emails 12 months after our last meaningful email exchange. I keep messages longer when needed for ongoing work or required records, and review those exceptions once a year.
The 12-month email period doesn’t apply to hosting, analytics, or radio records. Those records follow the relevant service arrangements and settings; the linked provider policies describe their practices. Plausible lets me delete the site’s analytics. An email address doesn’t let me reliably identify a particular visitor’s analytics records.
You can clear data saved by this site through your browser settings. Browser privacy tools may block analytics requests. Clearing browser data doesn’t erase email, downloaded files, or records already held by Cloudflare, Plausible, or SomaFM.
For questions or requests about information I hold or control, including access, correction, or deletion, email hello@ben-tan.com. Tell me which feature and information you mean. I’ll explain what I can identify and what action I can take. For information kept only in your browser, use the relevant feature controls or browser settings.
Links to LinkedIn, project sites, and other destinations lead to separate services with their own privacy notices.
This is an inbox, not a mausoleum with Wi-Fi, and I’d rather not torture myself by eternalizing our conversation.
4. Processing files and choices
Test Data Generator creates sample data and reads selected setup files in your browser. PII De-identifier processes selected files in browser memory. These tools don’t send inputs, file contents, or results to me or a processing server. Your browser creates output and setup downloads when requested. Loading the tools still makes the hosting and analytics requests described above.
Guide searches and choices stay in this tab’s memory. They aren’t uploaded or saved between visits. Reloading restarts the path. Copy puts the result on your clipboard when you ask; it isn’t secretly submitting your homework.
Saved records and downloaded files have their own controls below. Reload is not a universal eraser.
5. Saving work on your device
Some features keep browser records between visits. Reloading doesn’t clear them. Refresh has not been promoted to factory reset. Temporary tab memory and downloaded files are different again.
Study progress, saved notes and reflections, and separate drafts.
Saved rules, machine history, and CIPT records don’t expire on a timer. Each feature’s controls cover its own records; clearing this site’s browser data is broader. Your browser can also lose or remove stored data. Downloads and copies in another browser or at another address are outside that jurisdiction.
CIPT records are unencrypted, without a Studio sign-in. Other scripts on this portfolio’s origin can access them. Studio doesn’t upload these records to me or automatically sync across browsers or devices. Another route on this same origin isn’t a security perimeter.
CIPT backup files are readable and unencrypted. The .json extension is not a padlock. Progress JSON excludes separate drafts; draft recovery text is a separate file.
Import replaces progress; it doesn’t merge histories. If saving fails, your original saved progress and drafts survive. Optional tab-only recovery doesn’t save progress or draft changes. Reload returns the original saved copy. “For this tab only” is doing unusually honest work here.
Turning on the optional radio connects your browser directly to SomaFM. It receives your IP address and the request details needed to stream music. Turning it off stops the stream. It doesn’t delete information SomaFM already received.
Cloudflare delivers hosted files such as artwork and my resume. Tool outputs, setup files, and the machine’s record export are created locally when requested. Download clicks aren’t tracked in analytics. Clearing browser storage doesn’t remove files already downloaded; manage those wherever you saved them.
You can copy or download a CIPT voice-practice brief and take it elsewhere. Studio doesn’t use your microphone, record audio, call an AI or speech service, or upload transcripts. The conversation happens wherever you choose to have it.
Pages use temporary browser state to keep navigation working. EMA search text can travel in the page address, history, a copied link, or a request to the host. Plausible leaves it out of analytics. A link can carry more than your enthusiasm.
Guide choices restart on reload. PII De-identifier’s current file, result, generated codes and per-run random key stay in memory. Generator setups are downloaded files you reopen.
CIPT’s saved drafts survive reload. Clearing this site’s browser data can remove them; the file/rule example is not its saved study record.
:x ColdStorageRecords
Machine recordFirst + latest load; count; different can types
Saved ruleName; pattern; flags; optional column hint
CIPT recordsProgress + saved notes/reflections; separate drafts
The machine creates or updates its record on load if storage is available. A new record starts at one load; reloads count, and each distinct can type appears once. The record recognizes this browser and shows its history. There is no account, server copy, cross-device sync or analytics copy. Deleting it removes that local record; loading or restarting afterward creates a new one.
PII De-identifier reads saved rules at startup. Saving or deleting writes the updated list. The current file and result are not part of that saved list. Retired treatment presets may still leave old stored data; browser site-data clearing can remove it.
Answers and results, reported confidence and assistance, schedules, session and mock-exam history, notes, reflections, and separate drafts persist here.
CIPT drafts are separate from graded evidence: note/reflection drafts, unchecked selections, confidence/help choices, and lab state. Same-origin storage means the same scheme, host and port; another route is not a separate storage boundary. Browser profiles can also keep separate copies.
The machine uses localStorage key bending.records: firstSeen is the creation or reset time, lastSeen is the latest recorded load, visits counts loads, and cansVended lists distinct can types. Saved PII rules use redactorium.customRules.v1. CIPT also uses localStorage.
1 · Save replacementConfirm reset or reviewed import
2 · Clear draftsOnly after the primary save succeeds
If clearing failsKeep drafts; warn; pause draft editing
An import is replacement, not history merging. Invalid input or a failed primary save leaves the original saved progress and drafts intact. A successful primary save does not guarantee that draft clearing succeeded.
After a failed import save, a second confirmation can open the reviewed backup for this tab only. Both primary and draft storage writes pause. Changes remain in memory; export stays available. Reload restores the original saved copy.
To move between addresses or browsers: export at the old one, import and confirm at the new one, verify the result, and retain the original backup. This does not change the old copy.
Progress backup · JSONSaved progress, saved notes and manual reflections
Draft recovery · textDraft notes/reflections, unchecked choices and lab states
Both are readable, unencrypted downloads you request. The progress backup excludes the separate draft maps. The recovery text is for manual copying; it cannot be imported as progress.
Save notes/reflections before moving, or keep the draft recovery copy. Export progress and recover drafts before closing a tab-only session. Downloaded copies remain separate from browser site data.
:x ColdGuideDetails
Privacy Decision Guides runs search and decision paths in tab memory, including answer choices, selected US states, and optional checklist selections. It does not upload files, send answers to a processing server, or persist these choices in browser storage. Reloading restarts the path. Only the guide identifier enters the URL fragment; answers and results do not. Copy places the result on the clipboard when requested. The plain text includes the outcome, reasoning, next steps, answers, unresolved facts, relevant change and timing notes, source links, and recorded check or review status. If clipboard access is blocked, the same text is selectable. File downloads are not offered. Official source links open separately when selected.
:x ColdVoiceDetails
For voice practice, Studio can put a plain text brief on your clipboard or download it as a text file when requested. It contains displayed practice counts, selected topic prompts and rubrics, and source links, excluding personal notes and drafts. You choose any person or service outside Studio that receives the text you share. Studio does not access the microphone, record audio, make model or speech-service calls, or upload transcripts.
:x ColdNavigationDetails
The homepage/Bending Machine flight animation uses sessionStorage key bt.machine-arrival.v1 for the destination, animation mode, creation time, and position information. The destination consumes and removes it on arrival. Handoffs aged 30 seconds or more are rejected. This is separate from the machine’s record.
Atlas and Encyclopedia Mechanica (EMA) keep camera and simulation settings in page memory. EMA also records the selected entry, category, search text, and view in the address and browser history. A copied link or request for that address can include the search text. Plausible excludes query strings and fragments from analytics.
:x ColdStorageControls
Access and deletion controls
Enter the item number on the machine’s keypad and press #, labeled ENTER. Collect the can and scan its barcode.
51 · ACCESS
Shows the local record on an on-screen receipt. DOWNLOAD creates a JSON file in your browser without sending the record to a server.
52 · DO NOT SELL / SHARE
Prints confirmation that I do not sell personal information or share it for cross-site targeted advertising. You do not need to activate it to prevent those activities. It changes no settings and does not turn analytics off.
53 · DELETION
Prints a deletion request. Pressing DELETE on the actual machine’s receipt removes this browser’s machine record. Scanning the can alone does not delete it.
54 · PRIVACY NOTICE
Scan the can and choose GO to open this notice in a new tab.
These controls apply only to the machine’s local record. They do not retrieve or erase saved tool rules, downloaded files, analytics, hosting logs, email, or other service records. I do not receive this local record, so an email address does not let me retrieve it.
Identifier input: IP address + browser details. Raw values aren’t stored by Plausible; no visitor ID is saved in the browser.
Plausible uses request details to make a site-specific daily identifier. It doesn’t save a visitor ID in your browser.
Plausible doesn’t keep the raw IP address or full browser description. Reports can still include approximate location, browser and device type. Daily counting isn’t daily deletion.
Older treatment presetsCurrent tool doesn’t read/write them. Older stored data can remain until cleared.
This saved rule is an instruction: a name, a pattern and an optional column hint. It isn’t a copy of the file.
The current tool doesn’t read or write older treatment presets. Removing that feature didn’t automatically delete older stored data; browser site-data settings can remove it.
:x Warm Guides
Illustrated example1 / 4
/tools/privacy-decisions/#example-guideGuide identifier only · no answers
Example vendsAABCan types A/B are illustration labels
Reloads add visits. Repeated vends don’t add duplicate can types.
A new example record starts with first and latest load times, a visit count of one and an empty list of can types.
The machine’s record stays in this browser on this device: no account, server copy, cross-device sync or analytics copy. It has no automatic expiration.
:x Warm Controls
Illustrated example1 / 4
Example controls
Collect can → scan barcode
Local record exists
51 · ACCESSShow / download local recordAfter collection and scanning
Illustration only. No real browser record or analytics setting is changed.
After you collect and scan item 51, its receipt can show and download the machine’s local record.
These controls cover only the machine’s local record. They don’t erase rules, downloads, email or provider records.
:x Warm Deletion
Illustrated example1 / 4
Before deletion
Local browser record ▤
First load
Earlier visit
Latest load
Latest visit
Page-load count
8
Distinct can types
A, B
Four fields · illustrative values
Existing record
After a later return
Local browser record ▤
First load
11:00
Latest load
11:00
Page-load count
1
Distinct can types
None yet
Four fields · illustrative values
Saved rules · downloaded files · email · provider recordsSeparate from the machine’s local record
This is an example of an existing machine record, with a visit count and can history.
Deletion removes the old machine record; a later visit creates a new one. Rules, downloads, email and provider records are separate.
:x Warm Handoff
Illustrated example1 / 4
HomeMachine
Example A
Destination
Machine
Mode
Flight
Created
Departure time
Position
Animation position
Temporary handoff
Example age0 s
30 s or older → reject
Machine visit recordSeparate record, unaffected by consuming the handoff
The flight animation stores a temporary handoff: destination, mode, creation time and position.
The temporary handoff is separate from the machine’s visit record. Arrival removes it; an age of 30 seconds or more is rejected.
:x Warm Downloads
Illustrated example1 / 4
CloudflareHosted artwork / resume
Hosted file
● ● ●Your browser
Generated output
→
Your saved copies
Hosted copy
Generated copy
Wherever you save them
Browser site storage: separate from downloaded copies
Cloudflare delivers hosted files such as artwork and my resume.
Downloaded copies are managed separately from browser site storage. Download clicks aren’t tracked in analytics.
:x WarmStoragePlaces
Example: a current file, a saved custom rule, and a downloaded output in PII De-identifier. The controls below change this illustration only.
Illustrated example1 / 4
● ● ●Example browser
Working memory
Current file + result
Empty after reload
Browser storage
Saved custom rule
Rule removed
Downloaded copyOutside browser site storage
Saved file
Removed separately
The current file and result are in working memory. Saved rules are in browser storage. A downloaded copy is outside both.
Guide choices restart on reload. PII De-identifier’s current file, result, generated codes and per-run random key stay in memory. Generator setups are downloaded files you reopen.
CIPT’s saved drafts survive reload. Clearing this site’s browser data can remove them; the file/rule example is not its saved study record.
:x WarmStorageRecords
Machine recordFirst + latest load; count; different can types
Saved ruleName; pattern; flags; optional column hint
CIPT recordsProgress + saved notes/reflections; separate drafts
The machine creates or updates its record on load if storage is available. A new record starts at one load; reloads count, and each distinct can type appears once. The record recognizes this browser and shows its history. There is no account, server copy, cross-device sync or analytics copy. Deleting it removes that local record; loading or restarting afterward creates a new one.
PII De-identifier reads saved rules at startup. Saving or deleting writes the updated list. The current file and result are not part of that saved list. Retired treatment presets may still leave old stored data; browser site-data clearing can remove it.
Your answers and results, confidence ratings and help you report using, schedules, session and mock-exam history, notes, reflections, and separate drafts stay here.
CIPT drafts are separate from graded evidence: note/reflection drafts, unchecked selections, confidence/help choices, and lab state. Same-origin storage means the same scheme, host and port; another route is not a separate storage boundary. Browser profiles can also keep separate copies.
1 · Save replacementConfirm reset or reviewed import
2 · Clear draftsOnly after the primary save succeeds
If clearing failsKeep drafts; warn; pause draft editing
An import is replacement, not history merging. Invalid input or a failed primary save leaves the original saved progress and drafts intact. A successful primary save does not guarantee that draft clearing succeeded.
After a failed import save, a second confirmation can open the reviewed backup for this tab only. Both primary and draft storage writes pause. Changes remain in memory; export stays available. Reload restores the original saved copy.
To move between addresses or browsers: export at the old one, import and confirm at the new one, verify the result, and retain the original backup. This does not change the old copy.
Progress backup · JSONSaved progress, saved notes and manual reflections
Draft recovery · textDraft notes/reflections, unchecked choices and lab states
Both are readable, unencrypted downloads you request. The progress backup excludes the separate draft maps. The recovery text is for manual copying; it cannot be imported as progress.
Save notes/reflections before moving, or keep the draft recovery copy. Export progress and recover drafts before closing a tab-only session. Downloaded copies remain separate from browser site data.
:x WarmGuideDetails
Privacy Decision Guides keeps your search, answers, selected US states, and checklist choices in this tab’s memory. Reloading restarts the path. It doesn’t upload files, send your answers to a processing server, or save those choices in browser storage. Only the guide’s identifier goes into the page address after #, not your answers or results. Copy puts the result on your clipboard when you ask, as plain text. It includes the outcome, reasoning, next steps, your answers, unresolved facts, relevant change and timing notes, source links, and recorded check or review status. If clipboard access is blocked, the same text is selectable. The guide doesn’t offer file downloads. Official source links open separately when you choose them.
For voice practice, you can copy a plain text brief to your clipboard or download a text file. Both happen when you ask. It includes displayed practice counts, selected topic prompts and rubrics, and source links, without your personal notes or drafts. Any conversation is outside Studio, with whoever you choose to share the text with. Studio doesn’t access your microphone, record audio, call an AI or speech service, or upload transcripts.
:x WarmNavigationDetails
The flight animation between the homepage and Bending Machine leaves a temporary session-storage value with the destination, animation mode, creation time, and position information. The destination removes it on arrival. Handoffs 30 seconds old or older are rejected. This is separate from the machine’s record.
Atlas and Encyclopedia Mechanica (EMA) keep camera and simulation settings in page memory. EMA also puts the selected entry, category, search text, and view in the address and browser history. Copying a link can share that search text, and requesting the address can send it to the host. Plausible leaves it out of analytics.
Enter the item number on the machine’s keypad and press #, labeled ENTER. Collect the can and scan its barcode.
51 · ACCESS
Shows the local record on an on-screen receipt. DOWNLOAD creates a JSON file in your browser without sending the record to a server.
52 · DO NOT SELL / SHARE
Prints confirmation that I don’t sell personal information or share it for cross-site targeted advertising. You don’t need to activate it to prevent those activities. It changes no settings and doesn’t turn analytics off.
53 · DELETION
Prints a deletion request. Pressing DELETE on the actual machine’s receipt removes this browser’s machine record. Scanning the can alone doesn’t delete it.
54 · PRIVACY NOTICE
Scan the can and choose GO to open this notice in a new tab.
These controls apply only to the machine’s local record. They don’t retrieve or erase saved tool rules, downloaded files, analytics, hosting logs, email, or other service records. I don’t receive this local record, so an email address doesn’t let me retrieve it.
Identifier input: IP address + browser details. Raw values aren’t stored by Plausible; no visitor ID is saved in the browser.
Today’s visitor gets a daily identifier. There’s no tiny membership card tucked into your browser.
Plausible doesn’t keep the raw IP address or full browser description. Reports can still include approximate location, browser and device type. Daily counting isn’t daily deletion.
Older treatment presetsCurrent tool doesn’t read/write them. Older stored data can remain until cleared.
A saved rule is an instruction. Your original spreadsheet hasn’t been stapled to the back of it.
The current tool doesn’t read or write older treatment presets. Removing that feature didn’t automatically delete older stored data; browser site-data settings can remove it.
:x Hot Guides
Illustrated example1 / 4
/tools/privacy-decisions/#example-guideGuide identifier only · no answers
Example vendsAABCan types A/B are illustration labels
Reloads add visits. Repeated vends don’t add duplicate can types.
Four fields. I’ve supplied an entire vending machine to make them feel important.
The machine’s record stays in this browser on this device: no account, server copy, cross-device sync or analytics copy. It has no automatic expiration.
:x Hot Controls
Illustrated example1 / 4
Example controls
Collect can → scan barcode
Local record exists
51 · ACCESSShow / download local recordAfter collection and scanning
Illustration only. No real browser record or analytics setting is changed.
51 gives you access to the local record. I put the export behind a drink because apparently a plain button lacked ambition.
These controls cover only the machine’s local record. They don’t erase rules, downloads, email or provider records.
:x Hot Deletion
Illustrated example1 / 4
Before deletion
Local browser record ▤
First load
Earlier visit
Latest load
Latest visit
Page-load count
8
Distinct can types
A, B
Four fields · illustrative values
Existing record
After a later return
Local browser record ▤
First load
11:00
Latest load
11:00
Page-load count
1
Distinct can types
None yet
Four fields · illustrative values
Saved rules · downloaded files · email · provider recordsSeparate from the machine’s local record
An example history. It has been busy counting visits, not developing a right to remain forever.
Deletion removes the old machine record; a later visit creates a new one. Rules, downloads, email and provider records are separate.
:x Hot Handoff
Illustrated example1 / 4
HomeMachine
Example A
Destination
Machine
Mode
Flight
Created
Departure time
Position
Animation position
Temporary handoff
Example age0 s
30 s or older → reject
Machine visit recordSeparate record, unaffected by consuming the handoff
A perfectly good link is now hauling destination, mode, time, and position through session storage. I did this to move you between two pages.
The temporary handoff is separate from the machine’s visit record. Arrival removes it; an age of 30 seconds or more is rejected.
:x Hot Downloads
Illustrated example1 / 4
CloudflareHosted artwork / resume
Hosted file
● ● ●Your browser
Generated output
→
Your saved copies
Hosted copy
Generated copy
Wherever you save them
Browser site storage: separate from downloaded copies
A hosted file comes from Cloudflare. Even my resume has to be delivered from somewhere.
Downloaded copies are managed separately from browser site storage. Download clicks aren’t tracked in analytics.
:x HotStoragePlaces
One file, one saved rule, one downloaded copy. All from the de-identifier. “Saved” has been called in to explain itself. These controls only change the illustration.
Illustrated example1 / 4
● ● ●Example browser
Working memory
Current file + result
Empty after reload
Browser storage
Saved custom rule
Rule removed
Downloaded copyOutside browser site storage
Saved file
Removed separately
“Saved” is a spectacularly unhelpful word for three completely different places: working memory, browser storage, and your downloaded file.
Guide choices restart on reload. PII De-identifier’s current file, result, generated codes and per-run random key stay in memory. Generator setups are downloaded files you reopen.
CIPT’s saved drafts survive reload. Clearing this site’s browser data can remove them; the file/rule example is not its saved study record.
:x HotStorageRecords
Machine recordFirst + latest load; count; different can types
Saved ruleName; pattern; flags; optional column hint
CIPT recordsProgress + saved notes/reflections; separate drafts
The machine creates or updates its record on load if storage is available. A new record starts at one load; reloads count, and each distinct can type appears once. The record recognizes this browser and shows its history. There is no account, server copy, cross-device sync or analytics copy. Deleting it removes that local record; loading or restarting afterward creates a new one.
PII De-identifier reads saved rules at startup. Saving or deleting writes the updated list. The current file and result are not part of that saved list. Retired treatment presets may still leave old stored data; browser site-data clearing can remove it.
Answers and results, confidence and help you report using, schedules, session and mock-exam history, notes, reflections, and separate drafts stay in this browser.
CIPT drafts are separate from graded evidence: note/reflection drafts, unchecked selections, confidence/help choices, and lab state. Same-origin storage means the same scheme, host and port; another route is not a separate storage boundary. Browser profiles can also keep separate copies.
1 · Save replacementConfirm reset or reviewed import
2 · Clear draftsOnly after the primary save succeeds
If clearing failsKeep drafts; warn; pause draft editing
An import is replacement, not history merging. Invalid input or a failed primary save leaves the original saved progress and drafts intact. A successful primary save does not guarantee that draft clearing succeeded.
After a failed import save, a second confirmation can open the reviewed backup for this tab only. Both primary and draft storage writes pause. Changes remain in memory; export stays available. Reload restores the original saved copy.
To move, export at the old address, import and confirm at the new one, then check it and keep the original backup. The old copy does not receive a forwarding address.
Progress backup · JSONSaved progress, saved notes and manual reflections
Draft recovery · textDraft notes/reflections, unchecked choices and lab states
Both are readable, unencrypted downloads you request. The progress backup excludes the separate draft maps. The recovery text is for manual copying; it cannot be imported as progress.
Save notes/reflections before moving, or keep the draft recovery copy. Export progress and recover drafts before closing a tab-only session. Downloaded copies remain separate from browser site data.
:x HotGuideDetails
Privacy Decision Guides keeps your search, answers, selected US states, and checklist choices in this tab’s memory. Reloading restarts the path. It doesn’t upload files, send your answers to a processing server, or save those choices in browser storage. Only the guide’s identifier goes into the page address after #, not your answers or results. Copy puts the result on your clipboard when you ask, as plain text. It includes the outcome, reasoning, next steps, your answers, unresolved facts, relevant change and timing notes, source links, and recorded check or review status. If clipboard access is blocked, the same text is selectable. The guide doesn’t offer file downloads. Official source links open separately when you choose them.
For voice practice, you choose to put a plain text brief on your clipboard or download a text file, then take it elsewhere. It contains displayed practice counts, selected topic prompts and rubrics, and source links, without personal notes or drafts. Whoever you share it with gets that text. Studio doesn’t use your microphone, record audio, call an AI or speech service, or upload transcripts.
:x HotNavigationDetails
The flight animation between the homepage and Bending Machine leaves a temporary session-storage value with the destination, animation mode, creation time, and position information. The destination removes it on arrival. Handoffs 30 seconds old or older are rejected. This is separate from the machine’s record.
Atlas and Encyclopedia Mechanica (EMA) keep camera and simulation settings in page memory. EMA also puts the selected entry, category, search text, and view in the address and browser history. A copied link can carry that search text; a request for the address can send it to the host. Plausible leaves it out of analytics.
Enter the item number on the machine’s keypad and press #, labeled ENTER. Collect the can and scan its barcode.
51 · ACCESS
Shows the local record on an on-screen receipt. DOWNLOAD creates a JSON file in your browser without sending the record to a server.
52 · DO NOT SELL / SHARE
Prints confirmation that I don’t sell personal information or share it for cross-site targeted advertising. You don’t need to activate it to prevent those activities. It changes no settings and doesn’t turn analytics off.
53 · DELETION
Prints a deletion request. Pressing DELETE on the actual machine’s receipt removes this browser’s machine record. Scanning the can alone doesn’t delete it.
54 · PRIVACY NOTICE
Scan the can and choose GO to open this notice in a new tab.
These controls apply only to the machine’s local record. They don’t retrieve or erase saved tool rules, downloaded files, analytics, hosting logs, email, or other service records. I don’t receive this local record, so an email address doesn’t let me retrieve it.