Journals a reviewer can sign
Crypto books rarely fail at the debits and credits. They fail at the handoff.
Tax software has already done the inventory work. Koinly, Summ, or a prepared CSV knows the lots, the proceeds, the realised gain, and which rows were spam, failed, or internal. The general ledger wants something else: balanced journals, the chart the client already uses, and a trail a reviewer can tie to a source row.
Most firms still build that bridge in a workbook. Swaps arrive split across several lines. A missing cost gets filled so the journal will balance. A comment that says “payroll” overrides the hash. Unmapped names land in miscellaneous because the importer demanded an account.
The file balances. The explanation does not. Next year the workbook is gone.
Ledger Bridge is the desk for that job. Import the tax pack the firm already trusts. Map what has been decided. Park what has not. Export Xero or QuickBooks journals in which every line still points at the source row it came from.
Every figure ties. Every journal balances. No numbers invented.
What comes off the desk
Core principles enforced by the engine on every file.
Source stays the source
The tax file stays the source of cost, proceeds, and realised gain. If a cost is missing, the close does not invent one. Missing stays missing, and it shows up as a row to resolve.
Own transfers are not sales
Movement between the client’s own wallets and exchanges is not a sale. Staking deposits and withdrawals are not income; rewards still are. Spam and failed nil-value rows do not become journals.
Swaps as one conversion
A swap the tax file split into several rows is posted as one conversion, not as income on one side and a dangling expense on the other.
Evidence over comments
Unmapped counterparties and unconfirmed comments wait in suspense. They do not post to a guessed expense. Confirmed payroll hashes and named vendor maps code through, because someone already decided. A comment is a query, not an instruction.
The export balances, and the gain in the ledger is the gain in the tax file. An unexplained difference is a defect in the workpaper, not a footnote.
How a file actually runs
The five stages of a complete, verifiable crypto close.
Pick the source
Koinly needs the transactions CSV and the capital-gains CSV for that tax year — cost basis is not on the transactions file. Summ takes the report pack or the transactions CSV. Anything else comes in as a generic CSV. Files are parsed in the browser.
Bring the client's chart, or start from the crypto template
New accounts leave as their own chart import. Choose pooled digital assets or a ledger per token before export. Fiat stays in the bank accounts either way. Changing that choice after a year is posted is a restatement, not a setting.
Map vendors already agreed, and hash rules for attested batches
Everything else waits in the exception queue. That queue is the workpaper, and the list that goes back to the client.
Reconcile before anyone imports
If the bridge between the tax-file gain and the journals does not hold, the file is not ready.
Export the CSV
On Xero, use the classic journal importer. The newer manual-journals importer can collapse a file into one journal and date it at year-end — void that and reimport. Large files split so each part still balances. Then post, tie the digital-asset accounts to the workpaper, and file the exception log with the engagement.
What stays with the accountant
Clear separation between reconstruction mechanics and professional judgment.
Accounting policy & fair-value treatment
Ledger Bridge does not set accounting policy, and it does not turn a comment into payroll. Historical cost in the ledger and marked inventory in the tax file will differ, especially on yield-bearing tokens. That gap is expected unless the entity has adopted fair value. Say so in the workpaper.
Review hours vs reconstruction hours
The hours that disappear are reconstruction hours. What remains is review: the queue, the maps, the policy note, the import, the tie-out. Maps carry forward, so the next file does not start from a blank sheet.
Sold per file, not per seat
The close is sold per file, not per seat. Close is $490 for one company and one tax year. Practice is $390 a file, three prepaid, with maps reused and a workshop on the first file. Desk is $290 a file, ten prepaid. The deliverable is the CSV a reviewer has already tied — connecting Xero or QuickBooks to push it in directly is included too, at no extra charge.
“A crypto close is acceptable when the journals balance, the gain in the ledger is the gain in the tax file, and nothing was posted because a cell had to be filled. That is the standard. The spreadsheet was only ever a workaround.”
Entities, mapping rules, and journals are stored securely in your private workspace database so you can resume review on any device.