The reclassification layer your banking app is missing.

Every open banking integration ships with data quality debt. Spendaq closes the gap between what the API delivers and what your SMB customers actually see.

  Classification engine online
p99 latency 194ms
API version v1.4.2

Real-time transaction reclassification

Open banking APIs label transactions by raw MCC code or generic merchant name. Spendaq's ML classification layer corrects these labels against a taxonomy optimized for SMB expense reporting — Office Supplies, Business Travel, Software Subscriptions, and 80+ more categories. Batch or streaming, sub-200ms.

  • 80+ SMB-optimized expense categories
  • Confidence score per classification
  • Batch up to 500 transactions per request

90-day cash-flow forecast signal

Once transaction history is corrected, Spendaq generates a 90-day rolling cash-flow forecast signal. This feeds directly into your SMB banking dashboard — giving small business owners a reliable view of incoming and outgoing cash, not a misleading estimate from mislabeled data.

  • Rolling 90-day window, updated per-transaction
  • Signal includes: trend, variance, anomaly flags
  • Returns in same API response as reclassification

Event-driven delivery for live banking feeds

Don't poll. Subscribe to Spendaq's webhook events and get notified the moment a transaction is reclassified or a forecast signal changes. Designed for banking apps where latency and reliability are non-negotiable.

  • At-least-once delivery with idempotency keys
  • Signed payloads (HMAC-SHA256)
  • Exponential backoff retry with DLQ

Numbers that matter for banking products

94%+

Category accuracy on open banking feeds

<200ms

Median reclassification latency

80+

SMB-optimized expense categories

Works with your existing data feeds

No data pipeline re-engineering. Spendaq accepts transaction batches from any open banking source.

Plaid MX Technologies Tink Truelayer Finicity Yodlee Akoya Raw bank CSV Custom OFX/QFX

Start classifying in minutes.