Koinly Report Review

Grade your Koinly report — free

Keep Koinly. CryptoTaxEdge reads your export, grades coverage and accuracy, flags what needs a decision, and gives you your first 25 fixes free.

No account, no sign-up — the review runs right below. See plans on cryptotaxedge.com.

Your file is parsed in your browser — it is never uploaded or stored. Only public on-chain identifiers leave the page: transaction hashes, plus the wallet address for perspective. Verdicts on those public transactions may be cached to speed up repeat reviews — see the privacy policy.

1Export from Koinly

In Koinly, export your transactions as a CSV — the same file you already download at tax time.

2Get your grade

Every on-chain transaction is re-checked. Coverage, tag quality and open decisions roll into one grade, and every verdict shows its work.

3Download fixes

Corrected rows and your own decisions export as a changed-rows file that re-imports straight into Koinly.

Run the review

Upload your Koinly export. See what it missed — graded, with the exact rows to fix. Tags Koinly left blank get filled, every verdict shows its work, and your existing tags are never overwritten.

Step 1 · Upload & review

Parsed in your browser · only public transaction hashes and the wallet address are sent · your file is never uploaded or stored

Grey-area defaults applied to blank tags only — never overwrites Koinly: LP add/remove = taxable disposal · staking/rebase = ordinary income · wrap/unwrap = non-taxable 1:1 · liquid-staking mint = taxable disposal.

Change grey-area tax positions to your firm's standard

Contested treatments follow your firm's position. Defaults are the CryptoTaxEdge standard intake set. Set before running.

LP add / remove
Staking / rebase rewards
Wrap / unwrap
Liquid staking mint
Methodology

How grading works

The grade is an opinion about how much of your Koinly report would survive a second look — and it is deliberately conservative. Anything the review cannot stand behind stays visibly open, and open items pull the grade down until a person decides them. A flattering grade built on hidden question marks would be worthless, so there are none.

What the grade measures

Coverage. Koinly can only tag what it ingested — a transaction missing from the export entirely is a bigger reporting risk than a wrong tag. The completeness check compares the wallet's public on-chain history against your file, hash by hash; until it runs, the grade is marked provisional.

Tag quality. Every on-chain transaction in the file gets an independent classification, compared against the tag Koinly carries. Blank tags the review can fill with confidence count in your favor. Disagreements count against the grade until you decide them — the review never overwrites a tag you or Koinly chose.

Open decisions. Disagreements, engine-flagged transactions, low-confidence verdicts and rows with no verdict all count as open review items. Deciding them — in the queue above — is what raises the grade.

Where verdicts come from

Every verdict is produced the same way, in strict order of evidence strength:

  • Deterministic decode. What the transaction actually did, read directly from the chain. When the mechanics are unambiguous, no judgment is involved.
  • Verified rules. Classifications keyed to specific contracts and function signatures, verified before they are trusted.
  • Multi-source agreement. Independent readings of the same transaction are compared. Agreement raises confidence; disagreement lowers it — and is surfaced, never resolved silently.
  • Model judgment. Where the evidence above runs out, a model proposes a classification and states how sure it is. That verdict is labeled as judgment and treated accordingly.

One routing rule holds it together: a verdict below the confidence bar never becomes a tag. It goes to the decision queue with everything the engine saw, and you make the call. Genuinely contested treatments — the grey areas where practitioners take different positions — carry the position taken and the alternative, so your firm's policy decides, not a silent default.

Why there are no accuracy claims here

You will not find accuracy percentages on this page. Any such figure depends entirely on whose transactions are being measured — a wallet of simple transfers and a wallet of exotic positions produce very different results. The honest version is the one you generate yourself: upload your own export and grade what comes back.