From Kiosk to Wallet: Proving Which Payment Funded Which Transfer
Someone who believes they are in a relationship puts cash into a machine. Bitcoin moves to a wallet nobody can name. No single document says which payment funded which transfer. The link has to be built out of three separate systems, and the other side will work hard to pull it apart. United States v. Ian Freeman shows how that chain gets built, and where it tears.
What actually happened
Freeman, who describes himself as a New Hampshire radio host and church founder, began selling bitcoin in 2014 as part of what he called a mission to promote peace. On December 22, 2022, a jury in the District of New Hampshire, before Judge Joseph N. Laplante, convicted him of conspiracy to operate an unlicensed money transmitting business, operating one, conspiracy to commit money laundering, money laundering and tax evasion.
Then the district court took one count back, acquitting Freeman on the substantive money-laundering count: there was not sufficient evidence he knew the undercover agent went to the kiosk after Freeman said it was still there. The conspiracy-to-launder conviction stood.
On October 2, 2023, he was sentenced to 96 months, two years of supervised release and a $40,000 fine. On February 13, 2024, the Department of Justice announced he had been ordered to pay $3,502,708.69 in restitution to 29 victims, plus a preliminary forfeiture order. On July 29, 2025, the First Circuit affirmed (No. 23-1839), rejecting his argument that the money-transmitting statutes could not reach virtual currency.
The Department of Justice says he laundered over $10 million in proceeds of romance scams and other internet fraud, earned over $1 million in fees, and did not register with FinCEN.
The artifact: kiosk logs, trade history and bank records
Freeman sold through three channels, each leaving a different record. Bitcoin kiosks, machines where a customer deposits cash, scans a QR code tied to a wallet and receives bitcoin there. Ads and trades on localbitcoins.com. Direct negotiation over Telegram.
For localbitcoins and Telegram buyers he ran what he called a Know Your Customer procedure: a photo of a driver’s license, a photo of the buyer holding a handwritten note stating intent to buy, sometimes a phone number. None of that applied at the kiosks, which required no personal verification. (A dry aside: a handwritten note for one channel and nothing at all for the machines is a flexible reading of “Know Your Customer.”)
On the bank side, the Department of Justice says Freeman and his co-conspirators ran bank accounts in the names of four churches, among them the Church of the Invisible Hand and the Crypto Church of New Hampshire. (A name like that is, if nothing else, a form of disclosure.) It says he told customers, often scam victims themselves, to describe their deposits to the banks as church donations.
The records held a pattern no single transaction shows. In what the government calls “third-party trades,” Freeman took payment from one person and delivered the bitcoin to another. The opinion says he sent bitcoin paid for by people over 50 to recipients the evidence strongly suggested were scam artists, often after learning those payers sat far from the recipient and from others paying that same recipient.
What a software expert witness actually tests in a crypto-tracing case
Reconciling three systems. A kiosk record gives an amount, a time and a wallet address. A bank record gives a deposit by an account holder. The blockchain gives a wallet-to-wallet transfer, on a public ledger where each wallet is uniquely identifiable but says nothing about its owner, and where a sent transfer cannot be clawed back. Identity comes from the records around the chain, and whoever links them must show how.
Three clocks. A kiosk keeps local time. A bank distinguishes settlement date from value date. The blockchain has block time. A gap of minutes decides whether payment A funded transfer B. If the transfer looks earlier than the deposit, either the clocks are offset or the deposit did not fund it. The expert documents each offset and shows whether the match survives it.
Business record versus reconstruction. A platform’s trade history and a bank statement are generated in the ordinary course of business. A summary spreadsheet, or a profit figure assembled from scraped ads, is a reconstruction: a conclusion, with different weight and more places to attack.
The tax counts show the cost. Revenue agent Colleen Ranahan testified that Freeman filed no returns for 2016 through 2019. She calculated his annual localbitcoins profit from his advertisements and trade history, applied deductions, and computed taxes owed of $19,182.65, $66,033.55, $56,174.21 and $140,198.28 for those four years. On cross-examination she conceded she had used the standard deduction rather than itemized deductions for overhead, property tax and charitable giving, and agreed that had he itemized, he might have owed nothing. On appeal he did not dispute her method or her arithmetic. The First Circuit held the evidence sufficient: once the government identified unreported income and applied deductions from available information, the burden shifted to Freeman to produce rebutting evidence, and he never did. Cross-examination exposed the reconstruction’s limits. Only counter-evidence could have changed the result.
Where the chain breaks. A kiosk that captures no identity produces no record of who paid. A third-party trade severs payer from recipient. An undercover agent bought nearly $20,000 in bitcoin at a kiosk in a bar in Keene, New Hampshire; that episode sits behind the acquittal, where the missing link was knowledge, not a record. And a reconstruction rests on assumptions that can collapse under testing.
Walking each step and saying what each link does and does not hold is the work of a software expert witness.
Why this matters beyond one case
An Israeli litigator meets the same artifacts and reconciliation problems without needing American law. An opponent’s tracing report is often a reconstruction presented with the confidence of a business record. So the first question in any crypto-tracing claim is what the exhibit is: a record the system produced, or a person’s conclusion drawn from records. The second is which link rests on an assumption. That link gets tested, and where the data allows, an alternative goes on the table.
The above is general information and not legal advice. Case facts are drawn from the sources listed.