Legal
Privacy policy
Obsession is built so that your collection stays yours. This policy covers the Obsession apps for iPhone and Android, and this website, published by Big Helpers Software and Solutions Private Limited ("Big Helpers", "we").
The short version
Your collection is stored on your phone. Local phone use needs no account. Optional account sync and web access use an account; there is no advertising or advertising tracking. Optional online features send selected information for processing as described below; Android SDK diagnostics and server operational records are separate from your local collection.
The optional Obsession AI features process information for the feature you use, as described below. Feature availability and controls differ by platform and installed version. Manual features send requests when you ask; optional automatic look-ups and planned notification lines can send later requests under their enabled settings. Here is what those features send.
If you count the kilometres on your running or walking shoes (from version 1.5), the app reads your runs and walks from Apple Health or Health Connect. It never writes to them. Imported workouts are stored on your phone, outside iCloud; optional AI features can send selected workout details in the summaries described below. Here's how that works.
On Android, the sticker cut-out uses Google's ML Kit, which sends Google a small amount of diagnostic information. It never includes your photos, notes or prices.
Accounts and optional sync
Sync is optional. Your existing local or iCloud collection is not uploaded by signing in on this website. A compatible native app must show the migration choice and you must choose to sync that collection. Older app versions continue to use their existing local or iCloud storage. Use the same sign-in identity on each device; separate provider identities are not automatically merged by matching email addresses.
Sign-in information. When enabled, Google, Apple or taskss.in verifies your identity and shares an account identifier and available name, email address and email-verification status with Big Helpers. Your provider handles its own login. Obsession does not ask for or store that provider's password. Provider availability is shown on the sign-in screen.
What sync stores. Big Helpers stores the collection you choose to sync: shelves, possessions, names, descriptions, reasons, prices, purchase history, dates, loves, moments, saved places and precise locations, memoirs, badges, collection settings, and media such as photos, stickers and spin frames. Supported original purchase evidence includes JSON and HTML purchase records, text descriptions and the XLSX register. These files and collection text may contain personal information included in the original records or entered by you. Health workouts and mileage records are excluded from this account-sync contract and remain in their existing native storage. Server copies allow access through your signed-in iPhone, Android devices and web browser. The service stores revision and deletion records to prevent silent overwrites and to carry changes between devices. Uploaded images and documents require account authorisation; original HTML records are downloaded as files and never rendered inside the collection viewer. Collections are not public profiles.
Security and storage. Connections use HTTPS. The server authorises collection records and media against the signed-in account. Browser sessions use an essential Secure, HttpOnly cookie, and requests that change your account are protected against cross-site request forgery. The browser keeps a random device identifier for sync retries; it does not keep your provider tokens or collection in browser local storage. Authentication sessions, linked provider identifiers, app-device records and operational security logs are also processed to run and protect the service. When you create an account, the Big Helpers operator receives an internal email with your account details (such as email, sign-in method and device platform) so we can support and secure the service. Big Helpers processes the server copy to provide sync; this service does not provide end-to-end encryption.
Your choices. Sign out to end this browser session. Use Account & data in your collection to export the server's data manifest or request deletion. A manifest lists private image and document references; download their authenticated bytes separately for a complete offline archive. Deleting the phone app does not delete your server account. Optional AI processing remains separate from account sync, as explained below.
What the app stores on your device
- Photos you take or pick, and the stickers made from them
- The names, descriptions, reasons, prices and dates you enter, and why you love each thing
- Your loves: how many times you've loved each thing. Since version 1.1 the love heart replaces the 1 to 5 rating, and old ratings became loves.
- Your duel results and your settings, such as whether flashbacks and sounds are on
- Bragging rights write-ups and videos, nickname suggestions you asked for, and the nickname you wear
- The details you add to each thing (from version 1.6), such as its brand, size, fabric, stack height or kilometres, whether you know when you bought it, and the icon and counting word of a shelf you made yourself
- Your shoes' mileage (from version 1.5): which runs and walks went on which pair, and (from version 1.6) what else you wore on each one
The app stores this collection information in its private storage on your phone. Optional online features can also send the selected content described below for processing. Deleting the app removes its local storage; copies in device backups or exported files and records held by online processors are managed separately.
Camera, photos and motion
The app asks for the camera so you can photograph your things. It asks to add to your photo library only when you save a share card, collage or reel. To capture a spin, it reads the phone's motion sensors while you walk around an object. Photos you choose from your library are copied into the app; the app does not read the rest of your library.
Cutting out the background
Turning a photo into a sticker happens on your device. On iPhone this uses Apple's built-in Vision framework. On Android it uses Google's ML Kit subject segmentation, whose model is delivered and run by Google Play services on your phone; the photo is not uploaded.
Diagnostics sent by Google ML Kit (Android only)
ML Kit sends Google a small amount of diagnostic information so Google can keep the feature working: device and app details, an installation identifier, performance figures, the feature version and error codes. It never includes your photos, notes or prices. Google encrypts it in transit, uses it only for diagnostics and analytics, and doesn't pass it to third parties. See Google's ML Kit data disclosure and Google's privacy policy. We don't receive this information.
Fonts on Android
The Android app asks Google Play services for its two typefaces, Bricolage Grotesque and Figtree, through Android's downloadable fonts. Google Play services fetches them from Google if your phone doesn't have them yet. The request names only the fonts, nothing from your collection. Until they arrive, the app uses your phone's own fonts. The iPhone app uses the system fonts.
Sharing
When you share a card, collage or reel, the app hands that image or video to the app you choose (for example Messages, WhatsApp or Instagram) through your phone's share sheet. What happens next is governed by that app's own policy. Nothing is shared unless you tap share.
Obsession AI (optional)
The AI features are called Obsession AI. The app presents feature-specific disclosures and controls for optional processing. Manual features send a request when you ask; the automatic new-item lookup and notification-line settings described below can authorise later requests. Our server forwards the information needed for the selected feature to an outside AI service. Availability varies by platform and installed version. Settings › Obsession AI contains the available feature controls and master switch; Bragging rights and Lens have their own controls. Your nickname and the proud report send a request when you ask.
Bragging rights finds proud facts about one of your possessions from public reference pages, with their sources. When you ask, the app sends our server the possession's name and description, the name of its obsession, the exact product if you picked or typed one, and the app's language.
Automatic look-ups. Once you've agreed, the setting Look up new ones automatically is on. Each new possession you shelve is then looked up once, as soon as it has a name, and sends the same things. You can turn this off in Settings › Bragging rights.
Your nickname suggests five flattering nicknames. When you ask, the app sends our server the names of your obsessions, how many things are in each, the year each began, up to eight possession names for each obsession, your collector type and the app's language.
Lens (from version 1.3) names the thing in a photo you just took, so you don't have to type it. Only when you tap to let it, the app sends our server that one photo (made smaller), the names of your obsessions and the app's language. The photo is sent for that request and processed by our relay and the outside AI service. This statement does not establish the retention or logging practices of every service involved.
The proud report (from version 1.3) ends with a title and a line written by Obsession AI. When you ask for it, the app sends our server a summary worked out on your phone: how many possessions and obsessions you have, how long you've collected, the names of your most-loved, longest-kept and best-value possessions with their love count or days owned, the best value's cost a day (for example "1¢"), your most loyal obsession with its count and starting year, and your top three sticker colours.
Collection summaries and other answers. Collection totals and summary figures supplied by your phone are calculated there. Selected summary-generation paths check returned numbers against that supplied information. Web answers can include numbers from public sources, and entry answers can use numbers in the passages you send. These are distinct processing paths, not a promise that every number in an AI answer was calculated on your phone.
Did you know? (where available) shows a true fact about the things you love, with its source. The app sends our server the names of up to 12 of your obsessions, one possession's name for each, which facts you've already seen (as codes), and the app's language. The fact itself is general, not about you; the line about your own thing (for example how many days you've had it) is made on your phone.
Deep reads (where available) suggests a few long articles about your obsessions each week. The app sends our server the names of your obsessions and up to 20 possession names, and the app's language. No AI is involved in picking them. Before you read sends the article's address and title, and Obsession AI writes three points from the article itself. Articles open in your browser, under that website's own policy.
Post ideas (where available) writes a caption for a moment your phone found, such as your most-loved thing last month. When you ask, the app sends our server that possession's name and its obsession, the moment's numbers worked out on your phone (such as days owned, loves and years), your own "why" and "aha" lines for it if you wrote them, a fact you kept if you add one, the style you picked, the current caption when you tune it, and the app's language. Nothing is posted anywhere unless you share it.
The report (where available) gets a headline and up to three notes from Obsession AI. When you open it, the app sends our server the period and a summary worked out on your phone: what you spent in that period as one total (for example "$62"), how many new things and loves, the names of your best-value and most-loved possessions with their cost a day or love count, the names of new and sleeping possessions with the date each went quiet, how many of your top ten are older favourites, the rhythm of one obsession (how often, how many, since which year) and, if you planned one in the app, an upcoming trip's place and day.
Ask your collection (where available) answers questions like "What's my best value?". When you ask, the app sends our server your question, the names of your obsessions and up to 200 possession names, and the app's language. For numeric collection questions, Obsession AI can interpret the request for a local lookup and calculation. Answers from your written entries use the optional selected-passage processing described below. Look it up on the web (where available): if a question is about a thing itself rather than your collection (such as how to clean suede or when a camera was first made) and you tap Look it up on the web, the Android app sends Big Helpers the question you enter, the app's language and an installation identifier. This request builder does not automatically add collection names or numbers, but anything you include in your question is part of the submitted text. Our relay removes control characters and extra spacing, but does not automatically remove names, email addresses or other personal details. It forwards the cleaned question directly to its configured outside AI gateway, with grounding enabled to produce a sourced answer. This web handler does not itself write a local question or answer cache. Other reference and search features have separate result-cache paths; their existence does not mean that this web handler uses them. The gateway, search services, hosting records and backups have separate handling, and the absence of a cache write in this handler does not establish their retention or deletion. Do not include private information in a web-lookup question. The answer comes back with its sources, which open in an in-app browser.
Your running shoes (from version 1.5): if you count a pair's kilometres and Obsession AI is on, Obsession AI writes the pair's retirement plan, a line about a run and the words of your history. The app sends our server a summary worked out on your phone: the pair's name and what it is for, its kilometres used and left and a week, the date it may retire, run counts, average and best pace, easy-run heart rate then and now (by month), how many runs were wet, the names and kilometres left of your other pairs, and for one run its distance, time, pace, average and highest heart rate, cadence, weather words, first and last kilometre, and (from version 1.6) the names of the other things you wore with how many times each has been out. These summaries include details about individual workouts, such as heart rate, cadence and split information. They are not a complete Health record and do not include the workout route or a precise location.
Voice, receipts and answers from your entries (where available): a chosen recording is sent for transcription, and the transcript can be sent with collection names, the current date and currency to prepare an entry change. A receipt image you choose is sent to extract details such as price and date. Optional entry search sends text from your entries to produce numbers used in an index stored on your phone; selected matching passages can then accompany your question to answer it. The text can include possession names, descriptions, reasons, notes, journal moments and memoirs, with their relevant places and dates. Keeping the index on your phone does not mean that the source text never leaves it. From version 1.8 the index also holds short passages about your counted shoes (each pair's name, brand, type, carbon plate, stack, weight, tags, kilometres, run count and last run day) and a summary of each month of runs (pair names, kilometres, the longest run and race names); those passages are sent to produce their index numbers in the same way and can accompany a question. Local index controls can rebuild or forget that index. Information in recordings, images, questions and selected passages is processed by our relay and the outside service; their retention is separate from local storage and is not established by deleting a local recording or index. Proposed collection changes require your confirmation.
Big days (from version 1.8): when your phone finds a milestone, such as a pair passing 500 km, a pair's anniversary or your tenth possession, and Obsession AI is on, Obsession AI writes one line for it. The app sends our server the kind of big day and the facts your phone worked out for it: names (the pair, the shelf, the newest possession or a race) and numbers (such as kilometres, the pair's planned life, run counts, the month counting began and a weekly average), and the app's language. Once a week it may also send up to six short sentences your phone wrote about your own week (for example the weekday you run most, rainy runs, a streak or how you rotate pairs), with their numbers, so Obsession AI can pick one. If the line is busy or does not match those facts, your phone shows its own line. The list of big days your phone has found stays on your phone, outside backups and sync.
Captions for Make a share (from version 1.8): if you turn on the caption, the app sends our server the possession or shelf name, its obsession, the whole-number stats on the card (such as kilometres, runs, races, months, things, shelves and the year), the style and the app's language. The image is made on your phone and is not sent to us.
Notification lines (where available) writes the second line of the week's notifications. When your phone plans them, it sends our server, for up to eight notifications, the title it wrote, the possession and obsession names, the numbers in it, and your "why" line for that thing if you wrote one. The notification pictures are made on your phone.
Information supplied with a feature. Recordings, receipt images, questions and selected entry passages can contain prices, dates, reasons, places, journal text or other personal details you supplied. The descriptions above identify those processing paths; they are not a promise of automatic personal-detail removal. Lens and receipt features process the image chosen for that feature, and voice features process the recording you make for them. Each request also carries an app-generated installation identifier used for usage limits and app verification. Requests carrying it are processed with the selected content and normal network metadata. A random installation identifier does not make submitted content anonymous or guarantee that records cannot be linked.
Who processes it:
- Our server (Big Helpers, bighelpers.in) receives selected requests, forwards the information needed for the chosen feature to the outside AI service, and returns the answer. The relay also maintains usage counters under hashed installation and internet-address keys, application outcome logs and app-verification records. Hosting access and error records are separate: the configured access records include an internet address, request line, referrer and browser information. This does not establish that all other hosting records omit request contents. Request-triggered cleanup attempts to remove day counters older than two days and application logs older than 30 days; these are not guaranteed deletion deadlines. Hosting logs, backups, feature-result caches and outside processing can have separate retention. Their complete retention and deletion behavior is not established by those relay cleanup rules.
- Obsession AI, the outside AI service Big Helpers uses for the AI features of its apps, receives from our server only what each answer needs (the items above) and answers. The app's installation identifier and original internet address are not included in the selected AI request payload. This policy does not claim a verified retention, deletion or training configuration for the outside service.
- Wikipedia: reference features search for product, possession or obsession names and read matching public articles so answers can link to their sources. Those are names supplied by the feature and may contain personal details if you entered them that way. Requests come from our server. Each fact links to the page it came from; opening it uses your browser under Wikimedia's privacy policy.
- A web search service (where available): reference features may send query words, including possession or obsession names, to find public articles beyond Wikipedia. The inspected search request does not add the app's installation identifier or your original internet address, but query text itself can contain personal information. The separate web-lookup route described above forwards your cleaned question to its configured AI gateway with grounding enabled. These are distinct processing paths; neither is a guarantee about downstream retention or anonymity.
- Websites the search finds (where available): our server requests selected public pages to check facts and provide sources. Those sites receive a server request and its requested address and network metadata; this is not a guarantee that a selected URL or query conveys no personal information. When you open a link yourself, your browser contacts that site under its own policy.
Checking it's really the app: after you enable an Obsession AI feature, the app can use Apple's App Attest or Google's Play Integrity to help Big Helpers check that requests come from a genuine copy of Obsession. The service normally issues a signed session valid for one day. That expiry controls whether the session is accepted; it does not mean server records are deleted after one day. Verification can create challenge and usage-counter records associated with a salted hash of the installation identifier, and aggregate daily verification counts. On iPhone, the service also stores a key identifier, the public half of the verification key, an assertion counter, the verification environment and verification dates; the private key stays on the device. Creating or consuming challenges can replace or remove parts of a challenge record. The inspected verification code does not establish automatic expiry for every stored key or aggregate record, and uninstalling the app does not itself request their deletion. Hosting logs, backups and provider processing have separate handling. Google processes the Android check under Google's privacy policy.
Where the results are kept: write-ups, nicknames, Lens answers, proud lines and, where available, facts you keep, saved reads, captions, report words, answers and your notification inbox come back to your phone and are saved there with your collection. Saving results locally does not itself establish that every processing copy, cache or operational record elsewhere has been erased. You can clear saved write-ups in Settings, and take a nickname off at any time. The AI features page explains these features in more detail.
Running and walking shoes (Apple Health and Health Connect)
From version 1.5 you can count the kilometres on a pair of shoes, and from version 1.6 pairs on your Running shoes and Walking shoes shelves count them by default (you can switch this off for any pair). When you start, the app asks for permission to read your workouts from Apple Health on iPhone or Health Connect on Android. It only reads; it never writes anything to them, and never reads routes or your precise location. On iPhone, available workout details can include distance, time, pace and splits, heart rate, steps and cadence, climb, energy, and on some watches power and running form. On Android, version 1.6.1 (16) reads recorded running, treadmill-running, walking and hiking sessions, distance, heart rate, steps and cadence, and total and active energy when you grant the relevant permissions. Pace is calculated from distance and duration. This Android version does not request or read Speed, Elevation Gained or Power records. Previously saved workout details remain available; displaying old saved values does not request those permissions again. When both energy permissions are granted, the app may request both aggregates, prefers valid total energy for the saved outing, and uses active energy as a separately labelled alternative; the two are never added together. Android can also request older-history access separately, so a pair can count eligible workouts from more than 30 days before the initial grant, and optional background access to check for eligible workouts while the app is closed. These capabilities depend on platform support and the permissions you grant; background checks are periodic rather than immediate.
When a run or walk finishes, the app asks which pair you wore, and lists only your running pairs after a run and only your walking pairs after a walk. From version 1.7.1 the app keeps every run, walk or hike it reads, even when no pair counts it yet, and offers the pairs on your shoe shelves that are not counting yet; once, it looks back up to 400 days for runs no pair took. From version 1.8 a pair's type, weight, carbon plate and tags are details like any other and sync with your account if you use one; kilometres and runs never do. Your workouts and the mileage built from them are kept in a file on your phone that is not synced to iCloud or any server and is left out of your phone's backups. If the relevant Obsession AI writing features are enabled, the summaries described under Obsession AI can leave your phone. Android weather requests have separate mileage settings, as described next.
Weather (Android only). If you type a city in Mileage settings, your phone turns it into a position rounded to about 11 km and sends it to our server with the time a run ended (only for runs that ended in the last three hours). Our server asks the weather service of MET Norway (the Norwegian Meteorological Institute) for that place and hour, keeps the answer for an hour so others nearby get the same one, and sends back the temperature, humidity and conditions. The app also sends its installation identifier to our relay with this request; the relay receives normal network metadata such as the originating internet address. Weather is controlled by the mileage weather settings and can be requested separately from the AI-writing switches. On iPhone the weather comes from the workout's own information in Apple Health.
Paverun (optional, iPhone, from version 1.8)
Paverun is a running app from Big Helpers. If you connect it from Settings, Connected apps, Paverun asks you to agree, then hands Obsession a one-time code. Our server swaps that code with Paverun for a link key and keeps neither; the key is stored in this iPhone's Keychain, outside iCloud and backups. Obsession then talks to Paverun directly to keep your shoes in step: for each linked pair it sends the name, brand, type, tags, carbon plate and whether it is retired, and for each run you answer, which pairs you wore, identified by the run's Apple Health identifier. From Paverun it receives the same details for the shoes there and, for runs Paverun has, which pairs were worn with each run's date, distance and time. Kilometre totals, heart rate, routes and other Health data are not sent between the apps; each app works out kilometres from the same runs. Notes, prices and the rest of your shelves stay in Obsession. Disconnect deletes the key on this iPhone and tells Paverun to end the link; both apps keep what they already had. Paverun's own privacy policy covers what Paverun stores.
Reminders
If you turn on flashbacks, the app schedules local notifications on your phone. No notification server is involved. Where available, notifications get a picture made on your phone, and, if Notification lines is on, a second line written by Obsession AI as described above. They are still scheduled and shown by your phone alone.
Backups
Your phone's own backup (iCloud Backup or Google's Android backup) may include the app's data if you have backups turned on. Those backups belong to your Apple or Google account and are covered by their policies. We can't access them.
How to delete your data
Your local collection and saved results are stored on your phone. If you choose account sync, a server copy is retained for your account. Optional online features also involve processing and operational records described above. Removing local app data does not erase your server account or records held by online processors.
- Delete the app. This removes everything it stored on your phone: photos, stickers, notes, prices, dates, loves, write-ups and nicknames.
- Clear saved write-ups. In Settings › Bragging rights, tap Clear saved write-ups to remove every Bragging rights write-up from your phone. Turning off Look things up hides Bragging rights everywhere.
- Stop reading Apple Health or Health Connect. Turn off Obsession's access in the Health app (iPhone) or Health Connect (Android). Mileage settings can also stop asking after runs and walks, and a pair's menu has Don't count kilometres.
- Take your nickname off. Tap the profile button at the top of Home, open your nickname and tap Take it off.
- Device backups. Delete Obsession's data from your iCloud or Google account's backup settings. Files you saved or shared, such as share cards, collages and reels, stay wherever you put them until you delete them there.
The relay maintains daily usage counters under hashed installation and internet-address keys, application outcome logs and the app-verification records described above. Cleanup is triggered by requests: it attempts to remove day counters older than two days and application logs older than 30 days. These are cleanup thresholds, not a guaranteed deletion deadline, and they do not establish the retention period for hosting logs, backups or outside AI processing. For questions about these records, contact admin@bighelpers.in.
Delete an online account
In Your collection, open Account & data, verify your sign-in again if required, and type DELETE to request deletion. This immediately revokes account sessions and makes the account unavailable for normal use. It enters a 30-day recovery period. A one-time recovery code is shown; save it privately if you may want to restore the account. Recovery does not sign you in: you must then use your original provider.
Pending deletion is not permanent erasure. The current release retains the server collection during recovery and does not automatically purge it when that period ends. Contact admin@bighelpers.in for permanent erasure or help with a deletion request. Records required for security, legal obligations or backups may have separate retention. This request does not erase your local app collection, iCloud copies, exports or data held independently by sign-in providers and online processors. Manage those copies separately.
Children
Obsession is intended for adults and is not directed at children. The Android SDK diagnostics and the optional Obsession AI features described above still apply.
This website
Public pages load their pictures, fonts and videos from this site and run no advertising or analytics. Online access uses an essential session cookie. Choosing a sign-in provider redirects you to that provider, where its privacy policy applies. Private collection requests go to the Obsession service on this same website. Like any website, the server that hosts it keeps standard access logs (internet address, time, page requested and browser) to keep it running and secure. If you email us, we use your message and address only to reply.
Changes
We update this policy when storage, sign-in or online features change, describing what is processed and why before those changes are released. The date at the top shows when this policy last changed.
Contact
Questions about privacy: admin@bighelpers.in
Big Helpers Software and Solutions Private Limited