PaperPull

Collect years of statements and receipts. Skip the endless downloading.

PaperPull retrieves the statement and receipt history your banks, cards, stores and utilities already hold, organizes it as PDFs on your own computer or a server at home, and turns your purchases and transactions into spreadsheets. You sign in yourself, typing your password and answering the two-factor prompt in a real browser window. From there PaperPull only downloads. It never pays a bill, moves money or changes a setting (see how that is enforced). Free and open source, no account, no cloud. Made for archiving (into paperless-ngx or a folder of your own) instead of clicking through each site by hand.

On your computer, or on a server at home. The desktop app runs on Windows and macOS. PaperPull Server is the same app as one Docker image, for a NAS or home server beside Paperless-ngx. You use it from a browser on any computer in the house, sign in to each provider on a browser screen it shows you, and each run's new documents can go straight into Paperless. Both are free.

Watch it: a 77-second demo on the real panel with made-up accounts. Website: paperpull.net, with the getting-started guide, the provider directory, the Paperless-ngx tutorial and the FAQ (the most asked questions are also below). Download: the latest release, Windows 10 or 11 and macOS on Apple Silicon or Intel, Linux from source. Self-hosting takes the Docker image ghcr.io/rheeloaded/paperpull-server and three steps.

It also turns those PDFs into spreadsheets. Your receipts become one long table of every item you have ever bought, across every store and every account, with what it cost and when. The transactions inside your bank and card statements are read out of the PDFs into a workbook, one row each, and every statement is checked against its own printed balances, so a row you are looking at is a row that adds up. One button on the control panel, or one command. See Every purchase in one spreadsheet and The transactions inside your statements.

Runs on Windows and macOS (and Linux), with the same commands on each, and self-hosted on a NAS or home server with Docker.

PaperPull is free, and it costs money to make. There is a server and domains to keep paid, developer accounts for the signed Mac build and the Microsoft Store, the tools it is built with, and evenings on every provider to build it and to keep it working as sites change. If it saves you time, you can show your appreciation and support future development by sponsoring on GitHub or donating on Ko-fi, or by buying the Microsoft Store edition for $9.99, one time, the same program with a signed installer and automatic updates. Or download it free from the Releases page. Nothing is held back from the free build. See Support.

Seventy-two providers are supported today, all built on the same pattern. Seven more, SBA, Verizon Mobile, Meijer, American Family, Stripe, FedEx and Bank of America, are built and waiting for someone with an account to test them (#28, #31, #42, #45, #53, #54, #121):

App	Provider	Documents	Notes
affirm	Affirm	Loan agreements (Truth in Lending), one per loan	JSON API from inside the page, nothing clicked; no monthly statement exists for a pay-over-time account
adp	ADP Workforce Now	Pay statements, W-2s	Statement services from inside the page, nothing clicked but the tax statements' View statement that starts ADP's identity check, real Edge/Chrome. Requested in #46
aafmaa	AAFMAA (Armed Forces Mutual)	Annual statements, policy docs	ASP.NET WebForms; one documented disclosure dialog
ally	Ally Bank	Account statements, tax forms	JSON API; same-dated statements named from the PDF
alpaca	Alpaca	Brokerage and crypto statements, trade confirmations	Contributed by @JasperDale420, 22 documents saved on a full run. Each row's own Download pressed once, and only the download from the link the page hands that press is kept. A Paper account is never read
amazon	Amazon (any country's store, marketplace setting)	Order invoices (full history)	Per-year order pagination
amfam	American Family Insurance	Billing statements and any tax form listed with them	Untested, built without an account. Have one? Help test it Being tested by @jpfieber.
amex	American Express	Statements, Year-End Summary	Click-nav SPA; in-memory session
att	AT&T (Mobility, Fiber, Internet)	Monthly bills	Confirmed by @watling777 on two accounts, wireless and internet, with a full run on each (#26)
anthem	Anthem BCBS (Elevance, 14 Blue states)	EOBs, plan docs (all years), ID cards, letters	Health insurance (PHI); tRPC API, nothing clicked. Contributed by @riordan
apple	Apple (App Store and Apple Store)	App Store receipts for subscriptions, in-app purchases and paid apps, and Apple Store order invoices	Built and run on the maintainer's own account. Report a Problem's own purchase search from inside the page for the whole family, paid purchases only, each saved from Apple's own receipt, and the Apple Store's order list and invoices, nothing clicked, real Edge/Chrome
applecard	Apple Card and Apple Savings	Apple Card statements, Savings statements, tax forms	Confirmed by @watling777, whose Pilot and full run saved every statement and the 1099-INT forms. Requested in #52
bestbuy	Best Buy	Online, store-order and in-store receipts, and returns	Built on the maintainer's own account, 35 purchases back to 2015. The history through the page's own query a year at a time, each details page saved as the receipt, nothing clicked but the year menu, real Edge/Chrome, paced for Best Buy's bot protection
bilt	Bilt Card (serviced by Cardless)	Card statements	Contributed by @JasperDale420, 8 statements saved on a full run. Only the card's balance summary, View statements and a statement's own button are pressed, and the PDF is taken from the tab it opens
bofa	Bank of America	Checking, savings and credit card statements	Untested, built from a tester's two recordings, and its first two runs listed the statements and saved none. Have one? Help test it Statements & Documents' own list and PDF links, read from inside the page, with its Statements heading opened, and when the links have no address, one pressed to learn the page's own call for a PDF. Requested in #121
capitalone	Capital One	Bank and card statements, tax forms, letters	Ported by @davidrudnick; fresh live pilot pending
chase	Chase (credit cards)	Card statements	Real Edge/Chrome; per-card accordions + year picker
chasebusiness	Chase Business (business document center)	Business checking and business card statements, up to seven years	Contributed by @JasperDale420, 10 checking statements and the business card's first statement saved on full runs. One profile so far, so a debit account's statements are untested. Seven kinds of control pressed through the guard, real Edge/Chrome
citi	Citi (credit cards)	Monthly card statements	Statements API from inside the page, nothing clicked; the site lists about two years online
coinbase	Coinbase	Monthly account statements, Coinbase One Card statements, tax forms, gain and loss reports	Contributed by @JasperDale420, 66 statements and 2 tax documents saved on a full run. Each month's own PDF button on the Statements page pressed once and the download taken only from Coinbase's statements bucket, and tax PDFs from the list calls the Taxes page makes, nothing pressed. One Card statements from the Credit Card tab, each row's own PDF button pressed once, 13 saved on a full run on one account
costco	Costco Wholesale	Receipts, in-warehouse, gas, pharmacy and costco.com orders	Written from a member's recording and run against their account, 34 purchases back to July 2024 (#47). Each receipt names its part of Costco, Warehouse or Gas Station, for a file name pattern (#60). One membership so far, so a second Pilot is welcome. Two tabs, quarters walked, the warehouse receipt is a dialog rendered with printToPDF, real Edge/Chrome
discovercard	Discover (credit cards)	Card statements	Capital One is moving these accounts onto its own site. A card that has moved is read there by the capitalone app (#13)
dominion	Dominion Energy (VA)	Billing statements	Paginated MUI accordion; ~18-month limit
ebay	eBay	Order receipts, ten years of purchase history	Confirmed by @jpfieber, every order back to 2017. Order-details page rendered to PDF, nothing clicked but the order list's Show more, real Edge/Chrome. Requested in #44
lowes	Lowe's Home Improvement	Store and online receipts, and returns	Built on the maintainer's own account. The purchase history read by address, each purchase's details page saved as the receipt, nothing clicked, real Edge/Chrome
homedepot	The Home Depot	Order receipts, with model and store SKU numbers	Built on the maintainer's own account. The history through the page's own request, Home Depot's own print receipt taken without opening the print dialog, nothing clicked, real Edge/Chrome. Home Depot keeps two years online
etrade	E*TRADE	Statements, trade confirmations, tax forms	Confirmed by @watling777, whose Pilot and full run saved every statement (#36). No tax form seen yet
ezpassva	E-ZPass Virginia	Monthly and quarterly toll statements	Built on the maintainer's own account. The portal's own statements list read from inside the page, nothing clicked; about a year is online, so run it at least once a year
fairfaxwater	Fairfax Water (VA)	Water bills	Mendix portal, driven like a person; PDFs for the last year only, so run it quarterly
fedex	FedEx Billing Online	Shipping invoices	Untested, built from a login not connected to Billing Online. Use it? Help test it Requested in #54
fidelity	Fidelity Investments	Statements, trade confirmations, tax forms	Document Access Hub API from inside the page, nothing clicked; real Edge/Chrome
invest529	Invest529 (Virginia529 / Commonwealth Savers)	Official quarterly statements for every beneficiary and 1099-Q tax forms	Built and run on the maintainer's own account, 38 documents across three beneficiaries saved on a full run. Each read through the account's own API, nothing clicked
firsthorizon	First Horizon Bank	Checking statements, about two years online	Contributed by @JasperDale420, 24 statements saved on a full run. An account's panel and each statement's own Download PDF are the only presses. One checking account so far, so savings and card statements are untested
gap	Gap Inc. (Gap, Old Navy, Banana Republic, Athleta)	Order receipts	Lazy-loading history; ~13-month limit
irs	IRS (Individual Online Account)	Account, return, wage and income, and record of account transcripts	Contributed by @JasperDale420, 26 transcripts saved on a full run, untested by others. Have an IRS Online Account? Help test it Each transcript fetched by its own address, nothing clicked. Sign-in through ID.me is your own
kroger	Kroger (Pick 'n Save, Metro Market, Fred Meyer, Ralphs and the rest)	Receipts, in-store, fuel, pickup and delivery	Confirmed by @jpfieber on #41. Each receipt names its banner, Metro Market or Fred Meyer, for a file name pattern. Purchase-history API and receipt page, nothing clicked, real Edge/Chrome
github	GitHub	Payment receipts (Pro, Copilot, Actions, Sponsors and the rest)	Confirmed by @jpfieber on #43. Receipt links fetched or printed, nothing clicked
golden1	Golden 1 Credit Union	Statements, tax forms	Confirmed by @watling777, whose run saved every statement on both of his accounts, a credit card's named as one (#35)
ibkr	Interactive Brokers	Monthly Activity Statements	Contributed by @tylerverry, 60 of 61 months saved on a full run. The Statements page's own Activity Statement dialog, Monthly chosen and each month's own Download PDF button pressed once, and nothing else is clicked. Regional hosts .com, .co.uk and .ie
meijer	Meijer	Order receipts, in-store digital receipts where mPerks lists them	Untested, built without an account. Have one? Help test it Being tested by @jpfieber. Nothing clicked but its two tabs and an in-store receipt's own icon
myecp	MILITARY STAR (Exchange Credit Program)	Card statements	Built on the maintainer's own account, 25 statements back to 2023. The MyECP statements list and download from inside the page, nothing clicked
mypay	DFAS myPay	eRAS, CRSC, 1099-R, 1095	Government pay system; JSON API, nothing clicked
mtb	M&T Bank	Mortgage statements, escrow, 1098	Own online banking; you list, app expands all years
netbenefits	Fidelity NetBenefits (workplace 401(k))	Quarterly or monthly statements, made to order	The site generates statements on request; the app requests each period and renders it, nothing clicked
navyfederal	Navy Federal CU	Account statements	Per-account accordions; blob-tab PDFs
nelnet	Nelnet	Student loan statements, 1098-E, notices	Contributed by @tylerverry, 47 documents in one full run. The inbox list, paged and read from the page, each statement taken from its own Download button. The 1098-E is the Tax Info page printed after its own year's button is pressed
newrez	Newrez	Mortgage statements, escrow analysis, 1098	Confirmed by @watling777, whose run saved every document, three years of statements through the statements page's year picker and the 1098s (#38)
optum	Optum Bank (HSA)	HSA statements, 5498-SA and 1099-SA tax forms	Contributed by @JasperDale420, 21 statements and the 5498-SA saved on a full run. Every document a plain link on the member site, fetched by its own address, nothing clicked. One HSA so far, so more than one product is untested
paypal	PayPal	Monthly statements	Built on the maintainer's own account, 25 statements. The site's own statements list and download from inside the page, nothing clicked for a personal account, and three years are online
paylocity	Paylocity	Pay statements	Escher JSON API, enqueue-poll-fetch PDF; nothing clicked
pge	PG&E (Pacific Gas and Electric)	Billing statements	Salesforce portal with a paginated history. Contributed by @appchamp. Repaired in 0.37.1 and confirmed on a real account by @watling777 (#33)
portlandgeneral	Portland General Electric (Oregon)	Monthly bills, the Detailed Bill PDF	Built and run on a contributor's own account, 36 bills listed and the newest saved by a Pilot. The list and each PDF are asked of PGE's own account API from inside the signed-in page, nothing clicked. One account at a time
redcard	Target RedCard / Circle Card (TD Bank)	Billing statements	Statements table; per-year switcher
robinhood	Robinhood	Account statements, tax docs	"View More" pagination
sba	SBA (MySBA Loan Portal)	Loan statements, 1098	Untested, built without an account. Have an SBA loan? Help test it
schwab	Charles Schwab	Statements, tax forms, letters, trade confirmations	Ported by @davidrudnick; fresh live pilot pending
sdge	SDG&E (San Diego Gas & Electric)	Monthly bills, about two years online	Contributed by @JasperDale420, 8 bills saved on a full run. Each bill fetched by its own address from Billing History, nothing clicked. One account per run, the one selected on the site
shelbycounty	Shelby County Trustee (Memphis, TN)	Property tax receipts, one per payment, and the current tax notice while a balance is due	Contributed by @JasperDale420, 17 receipts across 16 tax years saved on a full run. A public lookup with no sign-in, the parcel ID goes in config.json. The page builds each PDF itself and the app keeps it, one guarded press per document
simmons	Simmons Bank (Banno online banking)	Loan statements, notices, tax forms	Contributed by @JasperDale420, 27 documents saved on a full run. Only a document row's own button is pressed, and the download taken only when it is that document's own. One loan so far, so a sign-in with a second account stops the list. A paid-off loan's documents stay online only until a date the bank names
smud	SMUD (Sacramento Municipal Utility District)	Monthly bills	Confirmed by @watling777, whose Pilot took the newest five and whose full run took the rest of the history (#34)
spectrum	Spectrum (Internet and Mobile)	Internet statements, about a year online, and Mobile statements printed from their pages	Contributed by @JasperDale420, 11 Internet statements saved on a full run. Mobile statements listed live, a whole Mobile run still to come. Each statement chosen by its own row. Two statements of one date, after an account transfer, are left for you to save by hand
stripe	Stripe (merchant Dashboard)	Fee invoices, tax forms such as the 1099-K	Untested, built on an account with no documents yet. Have Stripe invoices or tax forms? Help test it Requested in #53
statefarm	State Farm	Bills, renewal notices, ID cards, receipts, policy documents	Confirmed by @watling777, whose full run saved every document State Farm lists for his policies (#37)
target	Target	Receipts (Online + In-Store)	Print-capture
tmobile	T-Mobile	Bill statements	Bill-history page; detailed-bill download
treasurydirect	TreasuryDirect (savings bonds and Treasuries)	Current Holdings, Security and Payment History, saved from its pages, taxable years listed for you	Contributed by @JasperDale420, 3 documents saved on a full run, every PDF opened. 1099 pages unconfirmed. Have a taxable year? Help test it Its own links pressed in your tab, pages rendered to PDF
tsp	Thrift Savings Plan	Participant statements, 1099-R	Secure Mailbox API from inside the page, nothing clicked; downloading marks the message read
uber	Uber (rides and Uber Eats)	Ride receipts and Uber Eats order receipts	Built and run on the maintainer's own account, 24 receipts. The trip list and the Uber Eats order list from inside each signed-in page, Uber's own receipt PDF for each paid purchase, checked for its receipt ID, nothing clicked, real Edge/Chrome
ukg	UKG Pro / UltiPro	Pay statements	Per-employer tenant; JSON-API, nothing clicked
ups	UPS Billing Center	Shipping invoices	Built on the maintainer's own account. The Billing Center's own invoice list and download from inside the page, nothing clicked
usaa	USAA	Statements	JSON-API enumeration
usbank	U.S. Bank	Credit-card statements	Ported by @davidrudnick; fresh live pilot pending
vanguard	Vanguard	Account statements	Contributed by @tylerverry, 135 statements across five accounts. The statements page's own list, read while its year picker is walked, then each statement's own download control, pressed through the guard
venmo	Venmo	Monthly account statements, printed from the Statements page	Contributed by @JasperDale420, 24 months saved on a full run. Reads two years unless you set a From date, and says so when older months may hold more. Show more is the only press, and a month is printed only once every transaction is drawn
verizon	Verizon (Fios)	Bill statements	Real Edge (bot block); dropdown + CDP download
verizonmobile	Verizon Mobile (wireless)	Monthly bills	Untested, built without an account. Have one? Help test it
walmart	Walmart	Receipts	Hardened against bot detection
wellsfargo	Wells Fargo	Account statements, tax documents	Confirmed by @appchamp on a real account's credit card statements (#27). Checking, savings, mortgage and tax documents are untested. Have one? Help test it
wealthfront	Wealthfront	Statements, tax docs

⚠️ Read this first: these tools drive real, signed-in financial accounts. See SECURITY.md before you run or publish anything. In short: never commit your *-browser-profile/ folder, your config.json, or any downloaded PDF. The .gitignore blocks them, don't override it.

How it works (the shared design)

The one decision everything follows from

Your documents live on the provider's site, and it will only hand them to a browser that is already signed in. So PaperPull never tries to be you, it works beside you. You sign in yourself, in a real browser window, and the tool attaches to that window afterwards and reads.

flowchart TB
    you(["You"]) -->|"sign in · 2FA · device approval"| br["A real browser window<br/>its own profile · its own debugging port"]
    br -.->|"attaches over CDP, reads, never authenticates"| app
    subgraph app ["One app = one provider"]
        orch["Orchestrator<br/>discover → download → verify<br/>the same in every app"]
        site["provider_site.py<br/>selectors · URLs · download quirks"]
        core["paperpull-core<br/>naming · filing · state · CSV · PDF checks"]
        orch --> site
        orch --> core
    end
    app --> out[("Your folders<br/>PDFs + an index CSV")]

Loading

That single choice is why there is no password anywhere in this project, why 2FA and device approvals are never an obstacle, and why a provider tightening its login breaks nothing here.

In practice that first step is paperpull <app> login (or the app's own login.bat / login.command), which opens the browser for you, a plain Chromium for most apps, or your own installed Edge/Chrome for the few sites whose bot detection turns a fresh Chromium away (Walmart, Verizon, Chase). Each app gets its own profile and its own debugging port, so several signed-in browsers can sit open at once without colliding. Those profiles are separate from your everyday browser on purpose, so each starts with no extensions and no saved logins. They are real browser profiles, though, so a password manager installed into one, from the extension store in the sign-in window, stays there for every later sign-in to that provider. It is one profile per provider, so the extension goes in once for each provider you set up, not once for all of them.

Everything a provider knows lives in one file. provider_site.py holds every selector, URL and download quirk for that site. The orchestrator around it is the same in every app, and paperpull-core underneath it is shared. When a provider redesigns, the repair is one file, never a rewrite, and never a change to how documents get named, filed or tracked.

What one run actually does

flowchart TB
    D["Discover<br/>list what the provider still has"] --> Q{"Already downloaded?"}
    Q -->|yes| S["Skip it"]
    Q -->|no| DL["Download the PDF"]
    DL --> V{"Is it a real PDF?"}
    V -->|no| MR["Manual Review<br/>flagged, never silently lost"]
    V -->|yes| F["Classify, name, file<br/>+ append to the index CSV"]
    F --> OK["Mark downloaded_ok<br/>sticky, survives deletion"]

Loading

Three plain-text files carry the state, and you can read all of them:

File	Holds
discovery.json	what the provider showed us this run
progress.json	what happened to each document, including the sticky downloaded_ok
<Provider> Document Index.csv	one row per saved document, for humans and spreadsheets (receipt apps also keep an Order History.csv, one row per line item). Text that starts like a formula, an item a seller named =... for one, is written with a ' in front, so Excel opens it as text

That last step is what makes a re-run safe. downloaded_ok is keyed to the document, not to the file on disk, so you can import everything into paperless-ngx, delete the PDFs, and the next run still skips them. It only fetches what is genuinely new, and lists it in new-this-run.txt, which Discover, Verify, Rename and a dry run leave as the last download run wrote it. To fetch them again anyway, open the provider in the control panel, choose Restore deleted files under More and pick a year, or both a From and a To date, or add --redownload to a run from a terminal. Nothing is overwritten. A file still there stays as it is, and the app may save a second copy beside it.

Read-only by construction

Nothing that buys, sells, transfers, pays, deletes, or changes a setting is ever clicked, and all site interaction lives in provider_site.py where it can be read in one sitting. Every app carries a blocklist (FORBIDDEN_CONTROL_RE) of the words such a control would use. Most apps that press controls also decide which of the page's controls may be pressed with is_safe_control(), so a control must clear the blocklist and match a document allowlist (SAFE_DOC_CONTROL_RE), deny by default. The rest press only controls they find by a selector or by words written in the app, and the Target and Walmart apps check every receipt control against the blocklist, since a document allowlist would refuse the paging they press. Many apps click nothing at all during a run. They read the provider's pages by address, or the same data the provider's own page asks for. Each app's host allowlist refuses any stored URL that points elsewhere, and a repo-wide test checks every app's guard. SECURITY.md spells out which app does which.

One app, more than one person

In the control panel, add a person beside the Account box makes the second person's account, and the Account box then runs every action against it. From a terminal it is paperpull <app> add-account spouse, then paperpull <app> all --account spouse. Either way it gets its own profile, port and output folder beside the first one's, so no data mixes. Its port is the first free one from 9400 up, a range no provider comes with, so a provider you set up later never lands on it. Underneath it is a config.spouse.json beside the app's config.json, which the app also takes directly as --config, and the sign-in launcher takes the label too (login.bat spouse / ./login.command spouse). python tools/add_account.py spouse does every app at once.

PaperPull Server

PaperPull also runs on an always-on machine at home, a NAS or a home server with Docker, next to Paperless-ngx, and is used from a browser on any computer on your home network. The panel is the same, the providers' sign-in windows open on a browser screen the server shows you, and each run's new documents can be copied straight into Paperless's consume folder. You still sign in to each provider yourself, as on the desktop.

	Desktop app	PaperPull Server
Runs on	your Windows PC or Mac	a NAS or home server with Docker and an Intel or AMD processor
You use it from	its own window	a browser on any computer on your home network
Signing in	a browser window on your desktop	a browser screen the server shows in your browser
New documents	folders on your computer	folders on the server, and copied into Paperless-ngx's consume folder if you mount it
Who can reach it	only your computer	your home network, behind a password you choose
Updates	the next installer, or the Store	the new version in compose.yaml, then docker compose pull and docker compose up -d
Price	free, or the Store edition	free

Getting it running takes three steps, and SERVER.md has the rest, Paperless, a NAS's own Docker screens, updates and backups.

1. Copy server/compose.yaml and server/seccomp-chrome.json, or the two in PaperPull-Server-X.Y.Z.zip on the newest release, into a folder on the server, and set PUID, PGID, TZ and the data folder in compose.yaml.
2. Start it. Docker downloads the image every release publishes, ghcr.io/rheeloaded/paperpull-server.

   docker compose up -d

3. Open http://<the server's address>:8765 from a browser at home and choose a password, with the setup code the container prints in docker logs paperpull.

For HTTPS, the setup SERVER.md recommends, use compose.https.yaml with Caddyfile beside it in place of compose.yaml, see HTTPS.

It is made for a machine at home. A rented server works, but providers treat data-center addresses as likely bots and check far more often. It listens on your home network behind a password, so it has privacy and security pages of its own, PRIVACY-SERVER.md and SECURITY-SERVER.md. It is free, like the rest of PaperPull.

Quick start

These steps are for the desktop app. The server's are above.

Want to see it first? The panel's welcome screen offers Open the sample, a folder of invented statements and receipts. The Accounts page fills in and both spreadsheets build from it, so you can see exactly what PaperPull produces before you sign in to anything. Nothing downloads while you are in it, and Leave the sample puts you back. Every document in it says on its face that it is invented.

One-shot setup, one Python environment that the control panel and every app share, with the shared core in it.

setup-all.bat        REM Windows

./setup-all.command  # macOS / Linux

Then either drive everything from the GUI control panel, which lists every provider and person with what needs you, and gets new documents with one button per account:

gui\run_gui.bat

…or run a single app from the terminal, with one command for all of them (using amex as the example):

copy apps\amex\config.example.json apps\amex\config.json    REM then edit paths as needed
paperpull amex login            REM opens a browser, sign in yourself, leave it OPEN
paperpull amex pilot            REM download the newest few as a test
paperpull amex all              REM download everything available
paperpull amex resume           REM continue after an interruption
paperpull list                  REM every app it can see

paperpull is paperpull.bat on Windows and ./paperpull on macOS and Linux, or python paperpull.py anywhere. In the installed Mac app it is /Applications/PaperPull.app/Contents/Resources/paperpull, and in the installed Windows app it is paperpull.bat in the folder PaperPull was installed to, %LOCALAPPDATA%\PaperPull unless you chose another. Both work on the same installs the panel does. It finds the app by folder name, slug or provider, runs it on the right Python, and passes anything else straight through, so paperpull chase all --year 2025 --account spouse works. The commands are setup, login, discover, pilot, all, resume, verify, diagnose and dry-run. The panel names them for what they do. Sign in, Get new documents, Try the newest few first and Only check what's there sit on each provider's page, Continue appears after a run that stopped, and the rest are under More, each with a line saying what it does. A receipt provider's More also has Review names, for the receipts it was unsure how to name. Which dates (one year, or a date range) becomes the same --year, --start-date and --end-date every app takes, and Restore deleted files adds --redownload to one run scoped to a year or to both ends of a range.

Each app also has its own README with provider-specific details and quirks.

The one environment setup-all builds is gui/.venv. An app can still have an environment of its own, made by paperpull <app> setup or the app's own setup.bat / setup.command, and an app with one runs on it. setup-all used to make one in every app folder, about 400 MB each. To free that space in an older checkout, run setup-all again, then delete the .venv folder inside each app folder.

Knowing when to run it again

tools/status.py reads the state each app already keeps and reports how current every archive is. Copy it and status.bat next to your install folders and run it. It downloads nothing and changes nothing. The control panel's Accounts page shows the same, one row per provider and person, most in need first, each with the button that fixes it.

PROVIDER                     DOCS  NEWEST          AGE  ISSUES     STATUS
Some Payroll                   12  2026-06-18     72 d  2x month   !! OVERDUE
A Bank                         97  2026-06-30     60 d  monthly    *  due
A Mortgage                     93  2026-07-31     29 d  monthly       current
A Shop                        667  2026-08-13     16 d  -             ongoing

It reports the archives you have. Nobody holds an account with every provider, so a folder you never set up, or one left behind by a closed account, is left out entirely rather than listed as missing or overdue. An archive that was set up but never downloaded anything is called out by name, since that is the one state a single run fixes.

It answers "is something new probably waiting" rather than "when did I last run this", which are different questions. A run that only verified existing files still updates a timestamp while telling you nothing about whether a new statement exists. So the signal is the date of the newest document you actually hold, measured against how often that provider issues them.

The cadence comes from your own history and is measured per account, so nothing has to be configured, and a provider that moves from monthly to quarterly corrects itself. Measuring per account matters: one bank folder can cover several accounts, and pooling their dates makes a monthly cycle look weekly. Receipt archives are shown without a due date, because purchases arrive irregularly and "40 days overdue" would be noise.

It also looks for holes in the middle

Being up to date is not the same as being complete. An archive can hold a document from last week and still be missing whole years behind it, which is exactly what happened twice while building this: one mortgage archive held a single year of a seven year history, and a payroll archive quietly defaulted to year to date. Both looked healthy by their newest document.

So each series is also checked for periods that look missing from the middle:

Possible gaps. A run that looks current can still be missing
periods in the middle, so these are worth a look.
  A Bank
     2025-03-18 to 2025-05-17   1 missing   3 series, including Savings
     2026-01-01 to 2026-04-02   2 missing   3 series, including Savings

The hard part is not finding gaps, it is not inventing them. Plenty of real documents arrive irregularly, insurance ID cards and policy renewals among them, where a long quiet stretch means nothing was issued rather than something was missed. Flagging those would train you to ignore the report, so a series has to earn an opinion first: at least six documents, a median interval of ten days or more, and at least 65 percent of its intervals close to that median. Only then is an interval roughly twice the usual one reported, and it is reported as possible rather than certain.

Windows shared by several series are grouped, because one account missing a month is usually a quiet month, while the same window missing across several at once is what a run that failed part way looks like.

--html also writes a status.html dashboard you can bookmark, and --quiet prints only what needs attention. The dashboard reads no document contents and carries no amounts or account numbers, but it does list which providers you hold accounts with, so it belongs with your installs and is gitignored here.

Every purchase in one spreadsheet

The receipt apps (Amazon, Target, Walmart, Gap) record every line item they see while downloading, in an <Provider> Order History.csv beside the PDFs. tools/export_purchases.py gathers all of them into one workbook, one row per item across every provider and account, newest first, with an Orders sheet and a Summary of spend per provider per year. Amounts are numbers, so Excel can sum them. Nothing reads a PDF and it takes about a second.

python tools/export_purchases.py --root "C:\path\to\your\installs"

The panel has the same thing on its Spreadsheets page, one button, with a choice of one provider at a time (Amazon Purchases.xlsx). The file is rebuilt from scratch each time, so edit a copy, not the original.

The transactions inside your statements

A statement archive holds PDFs, and the transactions are inside them. tools/export_transactions.py opens each PDF a statement archive's index knows about, reads it line by line, and keeps the lines that have the shape of a transaction, a date, a description and an amount. Nothing in it is written for one bank. What makes it trustworthy is the statement itself. Every statement prints a beginning and an ending balance, and the transactions between them have to add up.

python tools/export_transactions.py --root "C:\path\to\your\installs"

Where the statement carries a running balance column, the sign of every amount is read off the balance and the statement reconciles to the cent by construction. Where it prints signed amounts, they are summed as printed and checked against every balance pair the statement shows, which is how a card statement that prints its summary three times is read right. A statement covering two accounts is reconciled one account at a time. A statement that does not add up is still exported, with the difference in the Statements sheet, so you know which rows to doubt. On the author's archive, every USAA and Navy Federal statement and 120 of 128 American Express statements reconcile. Brokerage statements never will, since their balances include market movement, and the sheet says so.

Amounts are the effect on the balance. Money in is positive, money out is negative, for a bank account and a card alike. Each PDF is read once and remembered in a cache beside the installs, so the first build of a big archive takes minutes and the next takes seconds. The panel's Spreadsheets page has this too, as Bank transactions, showing its progress.

Naming your files your way

Files are named 2026-03-09 Target Coffee Receipt.pdf unless you say otherwise. The panel's File names page shows the name your newest file would get, and builds it from parts you add, remove and move, the date in the shape you like, the provider, the order or document number, the account, the total and more, each shown with how many of your own files carry it. A part can be skipped when a document does not have it, so no name ends in a stray dash, and your newest files are shown under the new pattern as you change it. The pattern can still be written by hand under Edit as text.

A / in the pattern puts files in folders, inside the folder each kind of document already goes to. {provider}/{year}/{date:yyyy-mm-dd} {summary} files a Chase statement as Statements/Chase/2026/2026-03-09 Monthly Statement.pdf. A folder whose part is empty is left out, Manual Review stays one folder, and a pattern goes at most three folders deep.

One pattern covers every receipts app or every statements app, or any single app can have its own. After you save, the page offers to rename the files you already have to match, a preview first and then the rename, and nothing is downloaded again. Renaming moves files into the folders a pattern names, and back out again if you change your mind. The last run's new-this-run.txt is left as that run wrote it, so it names each file as the run saved it, before the rename. A line can then name a file that has another name now, or, when two files traded names, the other file, and the list's second line says so. The pattern language is in docs/file-naming.md.

A documents folder, for every provider or each person

Documents can go to a folder apart from the browser profiles, records and logs in your PaperPull folder, such as one Google Drive or another sync program keeps, one Paperless-ngx watches or a share on a NAS. In Settings, under Documents folder, give one folder for every provider, and each provider and person gets a folder of their own inside it, named as their PaperPull folder is, such as Chase Statements and Costco Receipts - spouse. Below it, pick a provider and person to give them a folder of their own instead, or keep their documents in the provider's own folder. One set that way keeps it when the folder for every provider changes, and Use the folder for every provider sets it back. The first save gives a folder in it to everyone whose config names no documents folder, one set back to the provider's own folder in an earlier version included, since that left no setting, so set those again below after it. The page lists each one that keeps a folder of their own, and Use it for every provider gives all of them their folder inside it at once, a copy of each config kept in the provider's Backups folder first.

Each download is still saved and checked in the provider's own folder, and one that fails its check is put aside in its Manual Review there. When a run from the panel ends, the documents it saved and checked move into the documents folder, into Statements, Tax Documents, Receipts and the other folders of each kind, with any folders your naming pattern adds. The provider's records, Manual Review, Logs, Diagnostics and Backups never move, so the documents folder only ever gets documents that passed their checks. The run's new-this-run.txt stays in the provider's own folder and lists what the run saved, each file where the run saved it, before the move.

Saving the setting moves nothing that is already there. Documents saved before stay where they are, and PaperPull still finds them, so nothing is downloaded again. The page says how many are still in the provider's folder and offers Preview move and then Move, which moves each one under the name it has, with its records following it, and puts everything back if a record cannot be written. Change the folder or set it back, and the same Move brings the documents to where it now points. A browser profile, a folder inside or around one, a folder holding a provider's own folder and a folder inside or around a person's own folder are refused. A provider or person added later gets their folder inside the folder for every provider as their first download run from the panel starts, and the run says so. One whose config already names a documents folder, their own folder included, is never changed by a run. The panel writes each one's folder into their config.json as "documents_dir", a full path, so a run from a terminal uses it too, and from a terminal the move is a Rename run with PAPERPULL_MOVE_DOCUMENTS=1 in its environment. On PaperPull Server the folder has to be inside the data folder, see SERVER.md.

Moving to another computer

The record of what each provider already downloaded is what keeps PaperPull from downloading a document twice, and it can move with you. On the old computer, press Export history in the panel's Settings. On the new one, or on PaperPull Server, choose that file under Moving from another computer? on the welcome page, or under Import a history file in Settings. Every provider and person in it is set up with their settings, and nothing they already downloaded is downloaded again. No sign-in or PDF travels in the file, so you sign in to each provider once on the new machine. It does list document titles, order numbers, totals and the account holders' names, so keep it as private as your documents. From a terminal, python tools/migrate.py --export and --import do the same for providers already set up.

Windows and macOS

Both have a package. Every release on the Releases page carries PaperPull-<version>-setup.exe for Windows, which installs the control panel, the shared core and every provider into your own user folder with no admin rights and no Python on the machine, and PaperPull-<version>.zip, the same folder for anyone who would rather not run an installer. The package is x64 and runs on Windows on ARM (a Snapdragon laptop, say) under the emulation Windows 11 provides, where the browser it drives is your own native Edge or Chrome and the Python side spends its life waiting on it. There is no native ARM64 build, because cryptography, the package PaperPull opens encrypted PDFs with, stopped publishing builds for Windows on ARM.

Two Windows editions, one program. The GitHub release is free. The Microsoft Store edition is the same package from the same build, for $9.99 one time, and what the price buys is convenience, a Store-signed package that installs with no warning, updates through the Store, and uninstalls cleanly, plus the knowledge that it keeps the project going. Nothing is held back from either edition. The AGPL permits selling copies and the source stays public. For macOS there is PaperPull-<version>-arm64.dmg for Apple Silicon and PaperPull-<version>-x86_64.dmg for Intel, both signed and notarized, so they open with no warning. About This Mac tells them apart, a Chip line for Apple Silicon and a Processor line for Intel. Drag PaperPull to Applications and double-click it. The Windows installer and the program inside it are code-signed with Azure Artifact Signing. SmartScreen can still show its prompt while the certificate is new, since it learns to trust a publisher from downloads over time. See Code signing policy below.

For a checkout of this repository, one download covers both. The double-click files each app keeps, and the one-shot setup, come in both flavors, and everything else is the same paperpull command on either:

Task	Windows	macOS / Linux
One-shot setup	setup-all.bat	./setup-all.command
Set up one app on its own	setup.bat	./setup.command
Sign in	login.bat	./login.command
Fix unsure receipt names	review_names.bat	./review_names.command
Test run	paperpull amex pilot	./paperpull amex pilot
Full run	paperpull amex all	./paperpull amex all
Control panel	gui\run_gui.bat	gui/run_gui.command

A second account is the same on both: paperpull amex all --account spouse.

review_names is in each receipt app, the ones that name a purchase from its items. It asks in a console for a better name for each receipt the app was unsure of, one at a time. The panel has the same review as Review names under a receipt provider's More, which lists those receipts and renames each one you type a better summary for, and that is how PaperPull Server offers it. An order saved as several invoices takes the new name on every one of them. The double-click file is also the one in a folder the installed app made, since the panel does the rest there, and in such a folder it runs on the app's own Python.

Only one thing genuinely differs. macOS keeps Playwright's browser inside an app bundle and in a different cache directory, and a couple of providers need a branded Edge/Chrome to get past their bot protection, that lookup lives in paperpull_core.browser and is handled for you.

Getting a checkout onto a Mac

To use PaperPull, the .dmg above is the way. This is for a checkout of the repository.

git clone is the smoothest route, it preserves the scripts' executable bit and macOS does not quarantine it.

If you download a release archive instead, prefer the .tar.gz: it keeps the executable bit, while a .zip drops it. After unpacking a download, macOS may also quarantine the scripts, so a double-click reports "cannot be opened because it is from an unidentified developer." Both are cleared in one go:

xattr -dr com.apple.quarantine .
chmod +x setup-all.command apps/*/*.command gui/*.command

Requirements

• Windows, macOS, or Linux, or for PaperPull Server, Docker on a machine with an Intel or AMD processor (see SERVER.md)
• For the Windows installer or the macOS .dmg, nothing else. They carry their own Python.
• For a checkout, Python 3.11+ and Playwright (installed by setup-all)
• A Chromium-family browser for signing in (Chrome, Edge, Brave, Vivaldi or Opera). Windows comes with Edge. From a terminal, the sign-in step can download a private Chromium instead, though a few providers need a real Edge or Chrome. Safari and Firefox cannot be driven this way.

FAQ

The questions people ask most, answered briefly. The full FAQ on the website has more, on where documents are kept, encryption, browser profiles, PaperPull Server's security, the spreadsheets and moving to a new computer.

Does PaperPull see or store my passwords?

No. Open sign-in window opens a real browser window on the provider's own sign-in page, and you type your password and answer the two-factor prompt there, the way you would on a new computer. PaperPull attaches to that window afterward and reads what you can see. There is no field anywhere in PaperPull for a provider's password, so there is nothing for it to store or send. PaperPull Server works the same way, through a sign-in window it shows in your browser, and asks only for the password to its own panel.

Can it move money, pay a bill or change my account?

No. It only downloads. Every app that clicks presses only controls it recognizes, and refuses any that would buy, pay, transfer or change a setting. Each app can visit only its own provider's addresses, and many apps click nothing at all. See Read-only by construction and SECURITY.md.

Does any of my information go to PaperPull?

No. There is no PaperPull account, no PaperPull cloud and no telemetry, so no crash reports, no usage statistics and no update check. The program contacts only the providers you sign in to, plus, once and only if you agree, a browser download when your computer has no usable browser. PaperPull Server also fetches Chrome from Google, and Chrome itself talks to Google as any Chrome does. See PRIVACY.md and PRIVACY-SERVER.md.

Will my bank notice, or lock my account?

PaperPull works in the browser you signed in to, reads only what your account can already see, pauses between documents the way a person would, and stops at every sign-in step and security check for you to handle. Even so, whether a provider's terms allow a tool like PaperPull is between you and that provider. A provider can limit or lock an account it believes an automated tool is using, and PaperPull cannot rule that out. TERMS.md says so too.

Does it get past CAPTCHAs or two-factor prompts?

No, and it never tries. When a site asks for a code, shows a robot check or signs you out partway, the run stops there, and everything it saved stays saved. Handle it in the sign-in window, then press Continue to pick up the rest. That human step is also why PaperPull does not run unattended.

What does it run on?

It runs in four places. There is no phone app.

• Windows 10 or 11. Windows on ARM runs it too, under the emulation Windows 11 provides.
• macOS, with one disk image for Apple Silicon and one for Intel. See Windows and macOS.
• Linux, from source. See Quick start.
• A NAS or home server, as PaperPull Server in Docker, on a machine with an Intel or AMD processor.

Is it really free? How is the Store edition different?

The free edition is the complete program, every provider and both spreadsheets, with no trial, no account and no ads. The Microsoft Store edition is the same package from the same build, for $9.99 one time. The price buys a Store-signed install with no warning and updates through the Store, and it supports the project. It unlocks nothing. See Support.

Which browser does it need?

One from the Chromium family (Edge, Chrome, Brave, Vivaldi or Opera). Windows comes with Edge. On a Mac, install one of them, since Safari and Firefox cannot be driven the way PaperPull needs. Each provider gets a fresh profile of its own, separate from your everyday browsing. If none is found, the panel says so. Run from a terminal, the sign-in step offers to download a private copy of Chromium instead, and asks first.

Will it download the same document twice?

No. PaperPull remembers every document it saved by what it is, not by where the file went. Move PDFs into your filing system, or let Paperless-ngx delete what it imports, and the next run still fetches only what is new. To fetch some again on purpose in the desktop app, use Restore deleted files under More.

Does it run on its own, on a schedule?

No. You start each run, and since many banks sign you out within minutes, a person stays in the loop by design. The panel's Accounts page says which accounts are due, and Get new for all runs each of them, one by one. An assisted monthly run is on the roadmap.

My provider is not listed. What can I do?

Request it. If you hold an account there, you are the ideal person to help, and testing a first version takes no programming (see Testing a provider). If you write code, Adding a provider shows how. See Contributing a provider.

What happens when a provider changes its website?

That provider may stop working until it is repaired. Each provider is a small app of its own, and everything it knows about a site lives in one file, so a redesign breaks only that provider and the repair is usually small. Make a support report from its More menu and report it. Documents you already saved stay saved.

Is a support report safe to attach to an issue?

Yes, it is built for that. Make a support report, under a provider's More menu, saves a file whose name starts with survey-. It holds counts and states and no text from your account, and it is the file to attach. Never attach a password, a browser profile folder, a token or an unredacted statement to anything you post.

How do I know the download is genuine?

The Windows installer and the Mac disk images are built by GitHub Actions from a tagged commit on a clean runner, never on a developer's machine, and the releases page shows each file's SHA-256 checksum. The Windows installer is code-signed, the Mac disk images are signed and notarized, and the Store edition is signed by the Store. Download only from the releases page, the Microsoft Store or paperpull.net. See Code signing policy.

How do I update?

PaperPull never checks for a new version of itself. GitHub's Watch button, set to releases, tells you when one comes out. On Windows, run the new installer over the old one, and the Store edition updates through the Store. On a Mac, drag the new version into Applications. On PaperPull Server, run docker compose pull and then docker compose up -d. Your documents and history are left as they are, and so are the desktop app's sign-ins. On the server, the restart can sign you out of a provider that keeps its session only while the browser runs.

Contributing a provider

No one has accounts everywhere, so PaperPull grows when people add the providers they use. If a bank, card, brokerage, utility, telecom, or retailer you use isn't here yet, you're the ideal person to add it:

• 📖 Adding a provider, a step-by-step guide (clone the closest app, rewrite one file, stay read-only, test, submit).
• 📋 PROVIDERS.md, what's supported and what's requested; claim one so nobody builds it twice.
• 📥 Can't build it yourself? Request a provider and someone with that account may pick it up.
• 🎬 Testing a provider, for people who have the account but do not write code. Sign in, press Record, click your way to one document, and send back what the app wrote down. That is usually all it takes to turn a scaffold into a working app.

Every contribution keeps the read-only, local, no-credentials design, see CONTRIBUTING.md and SECURITY.md.

Status & roadmap

• ✅ All seventy-nine apps pass their tests, the seventy-two that are supported and the seven still waiting for a tester, 32,425 of them across the repo. Most are in regular use by the author. The rest were built or confirmed by people who hold those accounts, named in the table above and in Thanks.
• ✅ Self-hosted. PaperPull Server, one Docker image published with every release since 0.43.0 as ghcr.io/rheeloaded/paperpull-server. See PaperPull Server.
• ✅ Packaged. A Windows installer and a signed, notarized macOS app, both built by GitHub Actions from the tagged commit, with checksums. The Windows builds are code-signed, and the Microsoft Store edition is live.
• 🔜 More providers: community-driven, see PROVIDERS.md.
• 🔜 Scheduled/assisted runs: a monthly "nudge + sweep" (e.g. the 1st) that opens the login browsers and then runs discover + resume across every app once you've signed in, delete-safe, so it only grabs what's new. Fully unattended runs stay out of scope by design: the tools never store credentials or bypass 2FA, so a human sign-in stays in the loop (long-session retailer apps may tolerate more automation than banks/cards).
• ✅ Shared core: the support code the apps used to duplicate now lives once in core/ as paperpull-core. An app declares an AppSpec, its folders, routing, CSV columns and config defaults, and keeps only its orchestrator and its *_site.py. tools/check_installs.py reports whether your installs have drifted from the repo.

Code signing policy

The Windows installer is built from a tagged commit of this repository by the Windows package workflow on a clean GitHub runner, never from a developer's machine, and the checksums of what it produced are published beside it. Signing goes through that same workflow, so a signed binary can only ever come from code that is in this repository or from the CPython archive whose SHA-256 it pins.

This program does not transfer any information to other networked systems unless specifically requested by the user. The only sites it contacts are the providers you sign in to yourself, and the only download it ever offers is a browser, at sign-in, with your agreement. The full statement is PRIVACY.md.

Team roles, current status and the full policy are in docs/code-signing.md.

Support

PaperPull is free and open source, and it costs real money and real time to develop. There is a server and domain names to keep paid, developer accounts for the signed and notarized Mac build and for the Microsoft Store listing, and the tools it is built with. Every one of the forty-odd providers took evenings to build, and each one needs repairing when its site changes, which they do.

If PaperPull saves you time, there are three ways to show your appreciation and support future development. All are optional, and none changes what the free build does.

• Sponsor on GitHub: github.com/sponsors/rheeloaded, one-time or monthly, from a card you already have on GitHub.
• Donate on Ko-fi: ko-fi.com/rheeloaded ☕
• Buy the Microsoft Store edition for $9.99, one time, the same program from the same build, with a signed installer that opens with no warning and updates through the Store. It unlocks nothing the free build lacks.

You can also help without spending anything: test a provider you hold an account with (see PROVIDERS.md), report what breaks, or contribute one. That is worth as much as a donation.

Thanks

PaperPull only reaches the providers people bring to it. These people built one, tested one against an account the author does not hold, or found a bug and diagnosed it to the line.

• @davidrudnick, the Capital One, U.S. Bank and Charles Schwab apps, and a long run of panel fixes.
• @riordan, the Anthem BCBS app, the first health insurer.
• @appchamp, the PG&E app, and the first run of the Wells Fargo app on a real account, which saved their credit card statements.
• @tylerverry, the Vanguard app, built and run against five accounts back to 2020, the Nelnet app, run against a student loan account back to 2019, and the Interactive Brokers app, run against five years of monthly statements.
• @JasperDale420, the Coinbase app, run against a retail account's monthly statements and tax documents, the Optum Bank app, built and run against an HSA, the Shelby County Trustee app, run against a Memphis parcel's sixteen years of receipts, the SDG&E app, run against a new account's first eight bills, the Alpaca app, run whole against a Live brokerage account, the First Horizon, Chase Business, TreasuryDirect, Venmo, IRS and Spectrum apps, each run on their own accounts, the Bilt Card app, run against a new card's first eight statements, the Simmons Bank app, run against a paid-off mortgage's statements, notices and tax forms, and NetBenefits for a sign-in with several plans.
• @marecabo, Amazon's legal invoice PDFs on the German store.
• @watling777, the tester behind AT&T, SMUD, Golden 1, E*TRADE, State Farm and Newrez, seven providers' worth of surveys and pilots in a weekend, and the PG&E pagination bug.
• @dertbv, three exporter and Navy Federal bugs, each diagnosed so exactly there was nothing to add, and the fix for Navy Federal Visa statements' closing date in 0.53.0 (#32).
• @liamrotheram and @OberstK, Amazon outside the United States.
• @jpfieber, who asked for eBay and became PaperPull's first sponsor before it was even built, then tested eBay, GitHub and Kroger to working, eBay through every order back to 2017, and is testing Meijer, American Family and Target.

If you tested a provider and are not here, say so on the issue and you will be.

Legal

This project is for personal archival of your own records. It is not affiliated with, endorsed by, or sponsored by any of the companies listed. All product names and trademarks are the property of their respective owners. Automating access to a website may be restricted by that site's Terms of Service, you are responsible for how you use these tools. Provided as-is, without warranty of any kind.

License. PaperPull is free software under the GNU Affero General Public License, version 3. You can run it, read it, change it and share it. If you distribute a changed version, or run one as a service for other people, the same license applies to what you distribute, source included. Contributions before 2026-09-18 were made under MIT and that permission is kept, see NOTICE.md.

Terms. Using PaperPull means accepting its Terms of Use, which the control panel shows the first time it opens. They say what PaperPull does and does not promise and limit liability, and take away nothing the license gives you.

Name. PaperPull is a trade name of TNE Enterprises LLC, which publishes this project. The name is not part of the license. A modified version needs its own name, so that anything called PaperPull is this project. What that allows and what it doesn't is in TRADEMARK.md.

Copyright (c) 2026 PaperPull, a trade name of TNE Enterprises LLC, and the PaperPull contributors.