Skip to content

Human+ML fraud analysis · comparison

ClearSale vs FeeGuard: analyst-reviewed fraud vs mechanical leaks

Prevention stops bad transactions before they settle. This page is about good transactions that settle correctly and then strand money on the way back out. ClearSale genuinely reduces the first category; nothing it does touches the second.

What ClearSale does

Brazil-rooted fraud prevention pairing statistical models with human analyst review. Source: https://www.clearsale.com.br.

That is real capability in its lane — and the lanes below show precisely where it ends.

Everything said here about ClearSale is summarised from its own public materials at the link above, never from private knowledge of how it works inside. Products change; if any line has gone out of date, tell us at feeguard@teamveristria.com and we will correct the page rather than argue about it.

Side by side: FeeGuard vs ClearSale

Eight capabilities a Connect platform needs. No empty cells — every row states what each product actually does.

CapabilityFeeGuardClearSale
Unreversed-transfer detectionYes — computed on every `charge.refunded` event: proportional expectation, rounding tolerance, ratio clampingNo — screening scope.
Application-fee leak detectionYes — application fees tested against reversals via Net Platform Margin; zero- and three-decimal currencies handledNo — fees untouched.
Dispute-loss clawback supportYes — recoverable figure computed on `dispute.closed` (transfer remainder, fee excluded); early warning on `created`; gated automationNo — disputes reduced indirectly.
Cross-border FX monitoringYes — realised rates measured against same-day currency-pair baselines; directional loss flagging beyond 0.8%No — fX untouched.
Runs on your live Stripe Connect account todayYes — read-only restricted key, roughly ten minutes to first findings, revocable anytimeDepends on deployment
Performs recovery actionsOne-click manual reversals (free tier); risk-gated automation with idempotency keys on Monitor ProVaries by product
Evidence & audit trailAppend-only trail on every action; finding-level evidence ids exportable as CSVVaries by product
Pricing modelDetection and the 90-day lookback free · monitoring is a subscription from $49/mo · success fee only when eligible funds are recoveredSee vendor site

Where ClearSale genuinely wins

  • — Human-review model catching patterns models miss.
  • — LATAM regional expertise.

Why platforms choose FeeGuard over ClearSale

  • After settlement, not before it. Prevention and post-settlement recovery never contend for the same event, so neither displaces the other.
  • All four vectors, natively. Unreversed transfers, kept fees, uncovered dispute losses and FX drag are separate detectors with separate arithmetic, not one heuristic wearing four labels.
  • Free before you pay anyone. Detection and the 90-day lookback cost nothing — measure your stranded dollars before committing to any vendor, including us.
  • Ten-minute, read-only setup. One restricted key you create and revoke; no migration, no seller re-onboarding, no payment-path changes.
  • Recovery is billed on results. The success fee applies only to dollars actually recovered and confirmed in Stripe — if nothing recovers, nothing bills.
  • Evidence-grade trail. Append-only logs with finding-level arithmetic — ready for controllers, auditors and diligence teams.

Bottom line

Human judgement catches bad actors. No analyst can flag a boolean nobody set. If your problem lives in human+ml fraud analysis, ClearSale deserves its reputation. If your problem is a platform balance lower than your ledger predicts — refunds succeeding while transfers never come home — that problem has exactly one purpose-built answer, and you are reading it. Run the free audit; the number settles the debate faster than any comparison page can.

Common questions

LATAM leakage angles?

Cross-border corridors add FX drag — measurement applies there as anywhere.

Is FeeGuard a better choice than ClearSale?

For human+ml fraud analysis, ClearSale is the specialist and we say so plainly. For detecting and recovering silent Stripe Connect losses — the four vectors above — FeeGuard is purpose-built and ClearSale does not perform that function. Most platforms comparing the two are feeling the second problem.

Can we run ClearSale and FeeGuard together?

Yes, commonly. They operate on different objects and different events; neither conflicts with the other’s workflow.

What does switching cost us?

Nothing upfront: read-only key, ten-minute setup, free detection. Recovery carries a success fee only on confirmed dollars — so trying FeeGuard risks no budget at all.