Sales Transactions

Create Transactions

Import sales transactions (e.g. wholesale sales) into Prediko so they feed into your demand and inventory pipelines.

Sales, not purchase orders

This endpoint records outbound sales — units sold to your customers. To record inbound purchase orders raised on your suppliers, use POST /orders instead.

The tenant is resolved automatically from your API key — you don’t need to specify it. The referenced SKUs and warehouses are validated and resolved by Prediko before the transactions are ingested.

Example template

Prefer a spreadsheet? Download the example Excel template — each row maps to one entry in the data array below (the field reference tab documents every column).

POST/api/v1/transactions

Query Parameters

Parameter	Type	Required	Description
store_name	string	Yes	Name of the store the transactions belong to, exactly as it appears in Prediko. Matching ignores case and surrounding whitespace; URL-encode names containing spaces.
store_id	string	No	Deprecated — internal store identifier. Still accepted, but use store_name. Supply one or the other, not both.

Request Body

Field	Type	Required	Description
data	array	Yes	Array of transaction line items (at least one).

Transaction Line Item

Field	Type	Required	Default	Description
sku	string	Yes	-	Exact SKU identifier the transaction applies to.
warehouse	string	Yes	-	Warehouse name the transaction was fulfilled from.
quantity	number	Yes	-	Number of units sold. Must be greater than 0.
unit_price	number	Yes	-	Price per unit. Must be 0 or greater.
timestamp	string	Yes	-	When the transaction occurred, as an ISO 8601 datetime (e.g. 2025-01-15T10:30:00Z). Include an offset such as Z or +05:30 to be explicit; a value without one is treated as UTC.
currency	string	No	-	ISO 4217 currency code for unit_price (e.g. USD, EUR).
order_id	string	No	-	External order identifier this line belongs to.
channel	string	No	-	Sales channel the transaction came through.
tags	string[]	No	-	Optional labels to attach to the transaction. Individual tags must not contain commas — tags are stored as a comma-separated list, so a comma inside a tag splits it in two.
discount_amount	number	No	0	Total discount applied to the line (absolute amount).

Example Request

cURL

curl -X POST "https://api.prediko.io/api/v1/transactions?store_name=Main%20Store" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "data": [
      {
        "sku": "SKU_ABC",
        "warehouse": "Main Warehouse",
        "quantity": 10,
        "unit_price": 19.99,
        "timestamp": "2025-01-15T10:30:00Z",
        "currency": "USD",
        "order_id": "WS-1001",
        "channel": "wholesale",
        "tags": ["b2b", "priority"],
        "discount_amount": 0
      }
    ]
  }'

Python

import requests

response = requests.post(
    "https://api.prediko.io/api/v1/transactions",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    params={"store_name": "Main Store"},
    json={
        "data": [
            {
                "sku": "SKU_ABC",
                "warehouse": "Main Warehouse",
                "quantity": 10,
                "unit_price": 19.99,
                "timestamp": "2025-01-15T10:30:00Z",
                "currency": "USD",
                "order_id": "WS-1001",
                "channel": "wholesale",
                "tags": ["b2b", "priority"],
                "discount_amount": 0,
            }
        ]
    },
)
result = response.json()

JavaScript

const response = await fetch(
  "https://api.prediko.io/api/v1/transactions?store_name=Main%20Store",
  {
    method: "POST",
    headers: {
      "Authorization": "Bearer YOUR_API_KEY",
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      data: [
        {
          sku: "SKU_ABC",
          warehouse: "Main Warehouse",
          quantity: 10,
          unit_price: 19.99,
          timestamp: "2025-01-15T10:30:00Z",
          currency: "USD",
          order_id: "WS-1001",
          channel: "wholesale",
          tags: ["b2b", "priority"],
          discount_amount: 0,
        },
      ],
    }),
  }
);
const result = await response.json();

Response

Returns the import result.

{
  "errors": []
}

Batch size

There is no fixed cap on the number of line items in data (the minimum is one). The whole batch is validated and resolved within the request, so very large payloads are bounded in practice by the request timeout rather than by a row count. Split bulk backfills into batches of a few thousand rows and send them sequentially — each batch is validated independently, so a failure only forces you to resend that batch.

Duplicates and retries

This endpoint is not idempotent. There is no idempotency key, and re-sending a payload submits those lines again rather than replacing the previous submission. Only retry a request that failed — if you receive a success response, treat the batch as submitted even if your client times out afterwards.

Supply an order_id on each line where you can. Lines are identified by order_id plus sku, and without one the identifier is derived from sku, warehouse and timestamp, so a stable order_id keeps a given sale identifiable across systems.

Processing

A success response means the batch was accepted — validation and SKU/warehouse resolution have completed, but ingestion finishes asynchronously. Transactions may take a short time to appear in Prediko and to be reflected in demand and inventory figures. The response carries no job handle, so there is nothing to poll: verify in Prediko rather than re-sending the batch.

Errors

Store resolution (422 Unprocessable Entity)

The store is resolved before any rows are read, so these errors mean nothing was imported:

Condition	detail
Neither store_name nor store_id supplied	store_name is required
Both store_name and store_id supplied	Provide either store_name or store_id, not both
store_name doesn’t match any store on your account	Unknown store_name '<name>'. Available stores: [...]
store_name matches more than one store	store_name '<name>' matches more than one store; use store_id instead
store_id doesn’t match any store on your account	Unknown store_id. Available stores: [...]

The Unknown store_name / Unknown store_id responses list the store names available on your account, so you can correct the value without a separate lookup:

{
  "detail": "Unknown store_name 'Main Stor'. Available stores: ['Main Store', 'EU Store']"
}

Only stores that are active on your account can be matched — a removed store is never resolvable.

Row validation (406 Not Acceptable)

If row validation fails, the endpoint returns the per-row errors reported by Prediko:

{
  "errors": [
    { "row": 1, "column": "sku", "reason": "RESOURCE_NOT_FOUND" }
  ]
}

For larger imports the response may also carry an errors_file_url — a downloadable file with the full error report, useful when there are too many failing rows to work through inline:

{
  "errors": [
    { "row": 1, "column": "sku", "reason": "RESOURCE_NOT_FOUND" }
  ],
  "errors_file_url": "https://..."
}

Resolution

Each line’s sku and warehouse must already exist for your account. Lines that reference an unknown SKU or warehouse are reported in errors and are not ingested.