EBCDIC encoder

An EBCDIC encoder translates everyday text into IBM mainframe code page 037, the classic US and Canada single-byte character set still found in COBOL copybooks, VSAM dumps, and MQ payloads. Paste plain text and choose to_ebcdic to receive space-separated hexadecimal EBCDIC bytes, or paste hex and choose to_ascii to recover readable characters.

How ASCII and EBCDIC code page 037 relate, and why unmapped characters matter

When the code point is higher, the character is omitted from the byte stream and recorded in unmapped with its string index and code_point so operators can strip or transliterate deliberately. Invalid hex tokens or an odd continuous length raise invalid_input rather than inventing padding. This landing targets searches for EBCDIC encoder, ASCII to EBCDIC converter, code page 037, and IBM mainframe character encoding—not packed decimal, not code page 500, and not UTF-8 byte dumps.

How to call the API and what the response contains

Send a JSON body with a required string field text (aliases input, value, and data) and an optional mode of to_ebcdic or to_ascii. The default mode is to_ebcdic: each mappable character becomes an EBCDIC CP037 byte written as lowercase hex, joined with spaces in result and ebcdic_hex. Mode to_ascii treats text as EBCDIC hex—pairs separated by spaces, commas, semicolons, pipes, or a continuous zero-to-nine A-to-F string whose length is even—and returns the recovered string in result and text. Input length is capped at one hundred thousand characters so a single request cannot exhaust Worker memory. On success both modes include mode, code_page set to the string 037, input, result, bytes as integer values zero to two hundred fifty-five, byte_count, char_count, mapped_count, unmapped_count, unmapped as an array of objects with index, char, and code_point, plus a short English message. Encode also exposes ebcdic_hex equal to result; decode normalizes ebcdic_hex to space-separated pairs even if you pasted continuous hex. Failed validation—missing text, non-string text, oversize input, an unknown mode, empty hex, odd-length continuous hex, or a token that is not exactly two hexadecimal digits—returns invalid_input and is never charged. Pricing is a flat $0.002 per successful request with unit item.

Practical tips, migration checks, and related tools

Round-trip checks are the fastest smoke test: encode Hello, confirm the first byte is C8 for H and that space would be 40, then decode the hex and confirm you recover the original string. Prefer this capability when the search intent is ASCII or Latin text versus IBM code page 037 hex; use a UTF-8 byte encoder when you need modern multi-byte sequences, and use Base64 tools when the goal is transport framing rather than character-set translation. When teaching, show the same letter as decimal ASCII, as CP037 hex, and as the wrong interpretation under ISO-8859-1 so students see why mainframe integrations still need an explicit EBCDIC step. Prefer to_ascii when you already copied hex from a dump, a CICS trace, or a hex editor and want the glyphs back. Deliberately include an emoji or a CJK character beside plain ASCII to assert that unmapped_count rises and that the EBCDIC stream contains only the mappable subset—migration scripts should fail closed when unmapped_count is non-zero if the target system cannot store those characters. Keep payloads under the published character limit; multi-megabyte binary datasets belong offline with iconv or dataset utilities. This tool is not a general EBCDIC family converter: code pages 500, 1047, and 1140 differ on currency and euro symbols, so do not assume the hex is portable across those CCSIDs. Failed tasks that return invalid_input do not consume prepaid balance, which keeps exploratory scripts cheap when learners paste incomplete hex groups.