Skip to main content
Bankbooks is Dataspike’s API for verifying Thai bank passbooks. Your user uploads a photo of a physical passbook; the API returns the account holder’s name (Thai and English), the account number and the bank, together with confidence signals and a flag for screenshots and scanned copies. One synchronous call, no applicant or verification object required. Use it wherever you need to confirm that a user owns the Thai bank account they give you — before a payout, a withdrawal, a loan disbursement or a claim payment — without manual review of passbook photos.

​
What you can do today

  • Extract the account holder’s name in Thai script, plus an English transliteration in account_name_en.
  • Extract the account number, digits and hyphens, tuned for Thai account formats.
  • Identify the bank as a stable enum (for example BANK_KRUNGTHAI), so you don’t need fuzzy matching on bank names.
  • Detect the form of the document: a paper passbook or an electronic statement from a banking app.
  • Flag screenshots and scans with is_scan, so photos of physical passbooks can be told apart from digital copies.
  • Route uncertain results to manual review using confidence_score and similarity_score, returned on every response.
Bankbooks is a standalone endpoint: each check is independent of applicants and verifications. To keep bankbook results on the same applicant record as KYC documents, AML status and liveness, use the Identity Suite integration — contact service@dataspike.io.

​
Who uses Bankbooks

​
Getting started

Bankbooks is one synchronous call, POST /api/v3/extract/bankbooks. Request parameters, the response schema and the error codes live in the Extract Data API reference.
1

Get an API key

Create an account in the Dataspike Dashboard and generate a key in the API section. Send it in the ds-api-token header on every request.
2

Test in Sandbox

Point your integration at the Sandbox environment and send a few real passbook photos, including bad ones: a card, a screenshot, a blurry shot. See Sandbox Mode.
3

Call Extract Data

Upload the image with issued_country=TH as described in the API reference. The response arrives in the same call — no webhook or polling.
4

Apply your decision rules

Route each result to approve, manual review or reject. Start from the recommended policy in Making a decision.
5

Go live

Switch to the Production environment and production key. Keep the check id from each response for audit and support requests.
A free test account is available for evaluation. Ask at service@dataspike.io.

​
Next steps

How it works

The pipeline, the signals and a recommended decision policy.

Coverage & FAQ

Supported banks, limitations, pricing and common questions.

API reference

Request parameters, response schema and error codes.