Batch Data Enrichment API

Each function has one or more named inputs; for most it is lookup. The catalog above lists every function's inputs. For JSON objects:

• A field named exactly like an input is used automatically (not case-sensitive): [{"lookup": "Cisco"}] needs no mapping. This works for several inputs at once, for example origin, lookup, and hscode for customs-duty-rates.
• Objects with a single field are used automatically by single-input functions.
• Otherwise, name the field: lookup_field=company. Other inputs use [input]_field.
• Give one value for every record with [input]_value, for example text_field=notes&to_value=Spanish with translate.
• All other fields, including nested objects, pass through unchanged. All records in a call must have the same shape.

Plain values (["Cisco", "Nvidia"]) need no mapping for single-input functions; row arrays and files use [input]_column. If a mapping is missing or ambiguous, the error message lists your record's fields and the exact parameter to add.

• JSON input returns one JSON document with Code, Function, RecordCount, ErrorCount, OutputFields, and Records, plus RequestedLanguage when a language was applied.
• JSON objects come back in the same order, each with the output fields appended after its own fields. If a record already has a field with an output field's name, the output is written as [name]_enriched, so none of your data is overwritten.
• Plain values and row arrays come back as {"Input": [...], "Output": {...}} per record.
• File input returns CSV or TSV when the job finishes: each record with the output fields appended as new columns, with an optional header row.
• Failed lookups leave that record's output fields empty and are counted in ErrorCount. See Limits and Billing for what happens after 20 failures.
• The X-Enrich-Language response header reports the language applied: a language code, ignored (the function does not take a language), or none.

Premium functions (marked in the catalog, plus custom) return every enriched value in English by default, or in any of these 16 languages with the language parameter, for the whole job, at the same price. Output field names stay the same in every language, so your downstream code does not change.

The value can be the two-letter code, the English name, or the native name, and is not case-sensitive. An unsupported language, or a function that does not take one, returns English and is never an error. The translate function is different: its target language is its own to input, for example to_value=Spanish. See international language support.

Enrich CRM records with firmographics. The company name is in the company field, and crm_id passes through.

Add NAICS and SIC codes with every description returned in German. The field is named lookup, so no mapping parameter is needed.

customs-duty-rates takes three inputs: origin, lookup (destination), and hscode. Fields named after the inputs map automatically; sku passes through.

With different field names, map them: origin_field=from_country&lookup_field=to_country&hscode_field=hs.

Translate the notes field of every record into Spanish: the text comes from a field, and the target language is one fixed value for the to input.

Describe the data with topic_value and list the fields to append with output_fields_value (a JSON array).

Enrich every company in column 1 of a CSV file with a header row, write a header row in the output, and save the results to a file.

Score the email addresses in column 3 of a CSV file, appending a 0 to 99 score and the reasoning to each record.

Without an API key, the same call returns HTTP 402 with the exact price for your records, and nothing is charged. An x402 client can pay it in USDC and repeat the call with the Payment-Signature header.

Returns HTTP/1.1 402 Payment Required with "amount": "20000" ($0.02 in USDC atomic units). The description states the price, for example 2 records at $0.01/record = $0.02 total.