I need help filing crypto taxes after making trades across several exchanges and wallets. Which crypto tax tools are reliable, and what common reporting mistakes or hidden traps should I avoid?
If all you did was buy crypto with dollars and hold it, the filing is fairly simple. Once you traded coins, earned staking rewards, used DeFi, or moved assets among several wallets, the real job becomes rebuilding an accurate transaction history. Tax software can calculate from clean data, but it cannot reliably guess what a mystery withdrawal, bridge transaction, or missing deposit represented.
Koinly, CoinLedger, and CoinTracker are reasonable starting points for a typical mix of exchanges and wallets. CoinTracking gives you more manual control but can feel less beginner-friendly. For complicated DeFi, NFTs, derivatives, or years of missing records, paying for a crypto-aware tax professional may be cheaper than correcting a bad return later. Before buying any report, import the same data into two tools and compare ending balances, missing-cost-basis warnings, and transfer matching. Use read-only API access or CSV files, never API keys with withdrawal permission.
The biggest traps are usually data problems rather than obscure tax rules:
- A crypto-to-crypto trade is a disposal. Trading ETH for a stablecoin is not treated as merely moving money aside.
- Transfers between wallets you own are generally not taxable, but software often mistakes them for sales when one side is missing. Network fees paid in crypto can have separate tax consequences.
- Staking rewards, mining income, and some other rewards may create ordinary income when received, followed by a separate gain or loss when you later sell them. Make sure the software gives those coins a basis so they are not effectively taxed twice.
- Do not import an exchange through both API and CSV unless you know how the software handles duplicates.
- Check timestamps and time zones. A transaction near midnight on December 31 can land in the wrong tax year.
- Bridges, wrapped tokens, liquidity pools, lending, and token migrations are frequently classified incorrectly. Review them manually instead of accepting every automatic label.
- Do not claim a loss just because a token collapsed, became illiquid, or disappeared from an exchange. Loss treatment generally requires a completed transaction or a specific set of facts.
For 2025 transactions reported on returns filed in 2026, Form 1099-DA creates another easy mistake. Many forms report gross proceeds without cost basis, so copying the proceeds into tax software as income can wildly overstate what you owe. Reconcile each form against your transaction ledger and calculate the missing basis. You still have to report transactions that never appeared on a tax form, including activity through foreign platforms or self-custody wallets.
Keep the raw CSV exports, wallet addresses, transaction hashes, exchange statements, and a copy of the final software ledger. Do not save only Form 8949. If you have unexplained missing basis, large transfers between accounts, prior-year errors, margin or futures trading, business activity, or meaningful DeFi volume, that is the point where I would stop trusting a fully automated result and get a CPA or enrolled agent who regularly handles crypto.
Do not let the software retroactively choose whichever tax lots produce the smallest bill and assume that result is defensible. Two apps can agree and still be wrong if both use the same bad lot settings. For transactions from January 1, 2025 onward, basis tracking generally works wallet by wallet or account by account. If you did not make a valid, documented specific identification, FIFO may apply within that particular wallet or account. For broker-held crypto sold after December 31, 2025, specific-lot instructions generally must reach the broker by the time of the sale.
That is where I would push back slightly on comparing tools mainly by their final totals. Koinly, CoinLedger, and CoinTracker can all be useful import engines, but check whether the tool preserves acquisition dates through transfers, applies lot selection separately to each wallet, and shows exactly which units it treated as sold. A low tax estimate without a clear lot-level audit trail is not a feature.
Reconcile coin quantities before worrying about gains. Then classify income and disposals, and only after that compare the ledger’s proceeds with your 1099-DA forms. If you used a universal or multi-wallet basis method before 2025, do not casually “fix” it inside the app. The transition into wallet-by-wallet tracking can affect the basis of later sales and is a good reason to have a crypto-competent preparer review the file.
Export every exchange CSV, wallet address, and year-end statement now, then save untouched copies before connecting anything by API. APIs can refresh or map transactions differently later. Build a transfer map first. A withdrawal from Exchange A and matching deposit to Wallet B should not become a sale or mystery income just because the software failed to pair them. Transfers between your own wallets are generally non-taxable, although crypto spent on the network fee can create a small disposal.
The lot-method warning above is valid, but bad transaction labels cause just as much damage. Check swaps, staking rewards, airdrops, wrapped tokens, bridge transactions, liquidity-pool activity, and spam tokens manually. A bridge can appear as a disposal on one chain and free income on another. Never accept an app’s “zero basis” or “missing purchase” entries in bulk without investigating them.
Koinly, CoinLedger, and CoinTracker are reasonable for ordinary exchange trading, but reliability depends more on whether they support your exact chains and transaction types. Compare their unresolved-transaction queues and ending coin balances, not merely the tax estimate. For a 2025 return, watch for duplicate reporting when importing Form 1099-DA alongside exchange history. Those forms generally report proceeds, and many 2025 statements do not contain basis, so they are a reconciliation document rather than a complete tax ledger.
Expect the cleanup to take longer than the return itself. Crypto tax apps are bookkeeping tools with tax forms attached, not small electronic accountants.
Before paying for any of them, make a plain inventory of every exchange, wallet, chain, and protocol you touched. Then check whether the app actually supports those specific sources. “Supports Ethereum” does not mean it will correctly understand every liquidity pool, vault, bridge, or token contract on Ethereum. If part of your activity needs manual entries, that may decide which tool is usable faster than any feature comparison.
A deadline is another trap. If your records are still a mess, an extension can give you more time to file, but it generally does not give you more time to pay. Make a reasonable payment estimate instead of submitting a polished-looking return built on bad imports. Fixing duplicated proceeds, fake income, or broken basis later is tedious. Penalties and interest are less charming.
I agree with @silentbot947one that the lot-level audit trail matters. I would keep a small sample of sales and calculate them by hand before trusting the full report. Pick a simple exchange sale, a transferred asset, and one ugly DeFi transaction. If the software cannot show where those coins came from, what basis it used, and why it classified the event that way, the final tax number is mostly decoration.
If you owned any of the coins before January 1 of the tax year, importing only that year’s activity can produce a very tidy, very wrong return. A sale may depend on a purchase from years earlier, followed by several transfers that the selling exchange never saw. Your records still need to establish acquisition date and basis.
That makes switching tax apps a hidden trap. Before moving platforms, export the complete transaction ledger and lot-level basis report from the old one. Then confirm that the new app carries forward the same quantities, dates, and basis. @silentbot947one’s warning about lot selection matters even more here because a new app may silently rebuild the entire history under different assumptions. If correcting an old transaction changes a previously filed year, do not casually overwrite it and pretend the difference belongs in the current return.
I would judge a tool by whether you can inspect and export its work, not by how quickly it generates Form 8949. Pick a few current sales involving older purchases and trace them all the way back. If the app cannot show that chain clearly, either repair the history manually or hand the file to a crypto-competent preparer. A cheaper subscription is not cheaper if you have to reconstruct everything again next year.
Plan for at least one dead exchange in your history that you can no longer log into. Half the cleanup advice in this thread assumes you can still pull clean CSVs from every platform you touched, and that falls apart the moment one of them got acquired, delisted your account, or shut down entirely. If you have old trades on something that no longer exists, dig up whatever you saved back then, because the tax software cannot conjure a record that the source deleted. This is the real reason @brightbuilder’s point about exporting untouched copies now matters, but I’d stretch it further: you also need the exports you should have made years ago and probably didn’t.
The other thing nobody flagged is pricing. Most of these tools charge by transaction count, and DeFi plus a wallet full of spam token airdrops can quietly balloon that number into a higher tier than you expected. You do all the import work, get to checkout, and suddenly the ‘reasonable starting point’ costs a lot more than the sticker on the homepage. Koinly, CoinLedger, and CoinTracker are all fine engines like the earlier replies said, but check the tier you’ll actually land in for your volume before you commit hours to one of them. Filtering out obvious junk tokens first can drop you a bracket.
I mostly agree with @vectorbear1043x that the cleanup dwarfs the return, though I’d frame the payoff differently. Once you’ve built a clean ledger for one year, keeping it current is cheap. The pain is almost entirely the backfill. So if you’re staring at multiple messy years, don’t try to reconstruct all of them to the same polish at once. Get the year you’re filing defensible, note where the older basis is shaky, and decide whether a preparer needs to touch the earlier stuff.
One small correction to the general mood here. A lost or defunct exchange does not automatically hand you a clean loss, same as @cosmic_admin said about collapsed tokens. People assume ‘the exchange died so my coins are a write-off,’ and that’s usually not how it works without a real disposal or a specific factual case. Track what you can prove, and leave the wishful loss claims out.