Ledgerly review: the sync works, the subscription trap does not
The best transaction categorisation I have tested, wrapped in a cancellation flow designed to waste your time.
- Published
- Last re-tested
- (8 days ago)
- Next review
- (in 4 months)
- Reading time
- 8 min
This review contains affiliate links
This review contains affiliate links. If you open an account through one, I may be paid a commission at no extra cost to you. The commission does not depend on the score, and the score was finalised before any link was placed.
Full disclosure policyLedgerly categorised 94.2% of 1,411 real transactions correctly with no manual rules. That is the highest figure I have recorded and it is the entire reason this product scores as well as it does. Against that: cancelling requires four screens, two confirmation dialogs and one 'are you sure' that defaults to keeping the subscription.
What works
- 94.2% first-pass categorisation accuracy on 1,411 real transactions
- Bank sync reconnected automatically after 3 of 4 institution-side breakages
- Genuine CSV and JSON export of all historical data, no paywall
- Shared household view does not require a second paid seat
What does not
- Cancellation is four screens deep with a dark-patterned default button
- Annual plan auto-renews with 3 days' notice, sent by email only
- Free trial requires card details up front and converts silently
- Support is email-only with a 31-hour median first response across 6 tickets
Buy it if
- People with messy multi-account finances who hate manual categorisation
- Households wanting one shared budget without paying twice
Skip it if
- Anyone who wants to try software without handing over a card
- People who need responsive support when a bank connection breaks
The number that matters
I connected five real accounts and let Ledgerly ingest 81 days of live transactions — 1,411 in total. I then reviewed every single one by hand against my own statements. It categorised 1,329 correctly on the first pass without any user-defined rules: 94.2%. For context, the same exercise on two competitors in 2025 produced 81% and 76%.
Where it fails
- Transfers between my own accounts were double-counted as income and expense 11 times until I set one rule
- Merchant names routed through a payment processor were categorised as 'Software' regardless of what was bought
- Refunds landed in the correct category but the wrong month in 4 cases
The cancellation flow, step by step
- Settings → Subscription (not visible from the account menu)
- 'Manage plan' → a screen offering a 40% discount to stay
- 'Continue cancelling' rendered as plain text next to a large 'Keep my plan' button
- A reason-for-leaving form with no skip option
- Final confirmation dialog where the default focused button is 'Go back'
Set a calendar reminder
Verdict
Situational, leaning positive. The categorisation engine is the best I have tested and it genuinely saves hours per month. The subscription mechanics are user-hostile in a way that is clearly deliberate. If you buy it, buy it monthly first, and diarise the cancellation date before you need it.
Regulated category — risk notice
Changelog
Every change to this review since first publication, including the ones that lowered the score. Corrections are logged here, never made silently.
Re-tested cancellation after Ledgerly's 4.2 release. Still four screens; the discount interstitial is new. Score lowered 7.1 → 6.9.
Re-ran categorisation on 300 fresh transactions: 93.7%, consistent with the original result.
First published after an 81-day live-data test.
Sources
Each source is recorded with the date it was read, because vendors change terms pages without notice.
- Ledgerly pricing page — accessed
- Ledgerly terms of service, s.7 (renewals) — accessed
- My categorisation audit spreadsheet (1,411 rows, anonymised) — accessed