Skip to main content

For developers

One API for transaction context.

Send a transaction to the REST API and get back merchant, category, labels and confidence as structured JSON. This page walks through the endpoint, the fields and the integration using a simplified example.

Request trial accessView field reference

At a glance

POST /v1/transactions/classify

Request

  • transaction_id
  • amount
  • currency
  • booking_date
  • counterparty
  • purpose

Response

  • transaction_id
  • merchant.name
  • category.id
  • category.name
  • labels
  • confidence

Simplified example of the TXNORA API.

API overview

A simple REST API. Plain JSON.

A transaction goes in as the request, structured context comes back as the response. That makes the API quick to integrate into existing backends.

  • REST endpoint

    Your backend posts the transaction as JSON and receives the result as a JSON response.

    POST /v1/transactions/classify

  • Clear matching

    The transaction_id from your request comes back unchanged in the response, so every result maps to its transaction.

    "transaction_id":"txn_demo_001"

  • Confidence per result

    A value between 0 and 1 indicates how certain the match is, so your product can decide what needs review.

    "confidence":0.984

  • Custom Endpoints

    For your own data structures, prompts and rules: custom endpoints process data to your specification, including large data volumes.

    Data structuresPromptsRules

Field reference

Every field, right in the example.

Type and meaning of every field in the example. Values come from the same sample transaction as the code.

Request fields

What your system sends in the example.

  • transaction_idType: stringYour own identifier for the transaction. It comes back unchanged in the response.Example"txn_demo_001"
  • amountType: stringSigned decimal amount, shown as a string in the example. Outgoing payments are negative.Example"-59.49"
  • currencyType: stringThree-letter currency code, for example EUR.Example"EUR"
  • booking_dateType: stringBooking date in YYYY-MM-DD format.Example"2026-10-05"
  • counterpartyType: stringThe counterparty as it appears on the statement, often truncated or inconsistent.Example"PAYPAL *ADOBE"
  • purposeType: stringPayment reference of the transaction as additional context.Example"ADOBE SOFTWARE"

Response fields

What the API returns.

  • transaction_idType: stringThe same identifier as in the request, so you can match the result.Example"txn_demo_001"
  • merchant.nameType: stringNormalized, recognizable merchant name.Example"Adobe"
  • category.idType: stringMachine-readable category key, for example for rules in your code.Example"software"
  • category.nameType: stringHuman-readable category name for display.Example"Software"
  • labelsType: string[]Contextual labels beyond the category.Example["SOFTWARE", "DIGITAL_SERVICE"]
  • confidenceType: numberValue between 0 and 1 indicating how certain the match is. The value shown comes from the example data and is not a measurement.Example0.984

POST /v1/transactions/classify

{
  "transaction_id": "txn_demo_001",
  "amount": "-59.49",
  "currency": "EUR",
  "booking_date": "2026-10-05",
  "counterparty": "PAYPAL *ADOBE",
  "purpose": "ADOBE SOFTWARE"
}

Simplified example of a TXNORA API request and response.

Integration

Integration in three steps.

Quick to integrate: your backend sends, TXNORA structures, your product decides.

  1. Send a transaction

    Your backend posts a transaction with amount, currency, booking date, counterparty and payment reference to the endpoint.

    POST /v1/transactions/classify

    "counterparty":"PAYPAL *ADOBE""amount":"-59.49""currency":"EUR"

  2. Receive structured context

    The response returns merchant, category, labels and a confidence value as JSON, ready for your logic.

    AdobeSoftwareConfidence98.4%
    SOFTWAREDIGITAL_SERVICE

  3. Decide in your product

    You decide what happens with each result. For example, route low-confidence results to manual review.

    Your example rule: review below 97.0%

    • AdobeConfidence 98.4%Result: Accept
    • REWEConfidence 99.1%Result: Accept
    • AmazonConfidence 96.2%Result: Review

Example from three typical transactions. The review threshold is an example rule that you set in your own product.

Operations & data protection

GDPR-compliant processing.

TXNORA processes your data on LLM endpoints hosted locally or in the EU.

  • No conventional public AI services

    Sensitive company data is not passed to conventional public AI services.

  • No LLM infrastructure of your own

    You integrate the REST API and send data. There are no models or endpoints for you to run yourself.

  • Scalable processing

    Processing scales with your data volume. For large data volumes and custom data models, there are custom endpoints.

Access

Trial access and documentation on request.

Send us a short request through the contact form. Within 24 hours, you receive free, time-limited trial access to the API, the developer documentation and our pricing.

Access
Free, time-limited trial access
Documentation
Included with trial access
Interface
REST API with JSON
Pricing
Usage-based and transparent
Request trial accessFrequently asked questions