The API push exporter delivers your transformed data directly to an external API. Instead of downloading a file and importing it into your system, WeTransform sends the output straight to the endpoint of your choice, as JSON requests whose shape you design yourself. It is the fastest way to feed an ERP, a PIM, a CRM, or any system with an HTTP API.
🔄 How it works
A file is submitted, manually or from an automatic source
The transformation runs: rules, validations, and corrections are applied
For each row of the output, WeTransform builds a JSON body from your target format's file settings and sends an HTTP request to your endpoint
A CSV report logs every request: status code, response, and errors
If you set a notification email on the exporter, a summary is sent with the report attached
⚠️ One request per row. If your output has 500 rows, WeTransform sends 500 requests to your endpoint. Use rate limiting (below) if your API has limits.
⚙️ Step 1 — Configure the destination
Open the Target formats page and click Exporters on your target format
Click Add a new exporter and choose API push
Give it a name, switch Enable on, and optionally set a Notification email for the summary reports
Configure the request: HTTP method (POST, PUT…) and the endpoint URL
Under advanced options, add your headers (for example
Authorization: Bearer …or an API access token) and cookies if your API needs them
🔒 Your endpoint URL, headers, and authentication tokens are stored encrypted at rest and only decrypted at execution time.
🧱 Step 2 — Design the JSON body
The shape of each request is defined on the target format itself: open General settings → File settings, pick JSON as the target file format, and click Start editing JSON body.
The JSON editor shows your structure as a tree of nodes on the left, objects, arrays, static values, and column references, with a live JSON Preview on the right showing exactly what will be sent. The JSON type (string, number, boolean) is inferred automatically from each column's data type.
You rarely need to build the tree by hand. Two shortcuts do most of the work:
📋 Paste a JSON sample — paste an example payload from your API documentation, and the whole node structure is created for you.
✨ Map columns with AI — the AI fills the variables of the pasted structure with the matching columns of your target format, automatically. Between the two buttons, a payload that would take an hour to wire by hand is ready in a minute.
For example, with columns sku, title, and price mapped into your API's expected shape, each row produces a request like:
{
"sku": "PROD-001",
"title": "Ergonomic Chair",
"price": 299.99,
"currency": "EUR"
}
💡 It works with GraphQL APIs too. The body can carry a query string and a variables object, so you can call a mutation like Shopify's productSet directly, with your columns injected into the variables.
In File settings, the toggle "Export all column values, even if they are empty" controls whether empty columns appear in the payload or are left out.
⏱️ Respect your API's rate limits
If your API enforces limits (for example 100 requests per minute), set the maximum number of requests per time window in the exporter's configuration. WeTransform throttles automatically: when the limit is reached, requests pause and resume once the window allows. No rows are dropped, they are simply delayed.
📊 Reports and notifications
After every push, a CSV report is generated with one line per request: row number, request body, response status code, response body, and error message if any. The report is available for download for 30 days from the submission details page. With a notification email set on the exporter, you receive a summary after each push with the report attached: rows processed, requests sent, success rate.
👨💻 For your developers: the full technical reference (node types, type inference, HTTP settings) is at api.wetransform.com.
👉 What to do next