Float to IEEE-754 converter

An IEEE-754 float converter shows the exact bit pattern a computer uses for a floating-point number. Non-numeric decimal text is rejected with invalid_input instead of being silently coerced.

What IEEE-754 binary32 and binary64 store

The same layout encodes special values without a separate tag: an all-ones exponent with a zero significand is signed infinity; an all-ones exponent with a nonzero significand is a NaN; an all-zero exponent with a zero significand is signed zero; an all-zero exponent with a nonzero significand is a subnormal number that fills the gap near zero. Engineers debugging serialization, teachers explaining floating-point error, and reverse engineers reading hex dumps all need to see that layout explicitly. IEEE-754 is the interchange format almost every modern CPU, GPU, and language runtime uses for binary floating-point arithmetic. A binary32 value occupies thirty-two bits: one sign bit, eight exponent bits with bias one hundred twenty-seven, and twenty-three explicit significand bits with an implicit leading one for normal numbers. A binary64 value occupies sixty-four bits: one sign bit, eleven exponent bits with bias one thousand twenty-three, and fifty-two significand bits. This capability does not invent a custom encoding: it uses the platform DataView big-endian float getters and setters so the hex string you get matches the conventional memory dump of a big-endian IEEE word, and the binary string is simply those bytes expanded bit by bit from most significant to least significant.

How to call the API and what comes back

Send a JSON body with at least one source field. To encode a decimal, set value (aliases decimal, number, n) to a finite number or to Infinity, minus Infinity, or NaN when those specials are representable in your client. To decode a bit pattern, set hex to an optional-0x hexadecimal string of up to eight digits for binary32 or sixteen for binary64, or set binary (alias bits) to a string of zeros and ones of matching length; shorter patterns are left-padded with zeros. Optional precision (aliases format, width) selects thirty-two or sixty-four bit width and accepts names such as single, double, binary32, and binary64; the default is sixty-four. Optional direction (alias mode) forces from_decimal, from_hex, or from_binary; auto picks from the fields present and prefers value when several are set. On success the response includes value (with string forms for NaN and infinities), is_neg_zero, precision, format name, bits, hex, hex_prefixed, binary, sign_bit, exponent_bits, mantissa_bits and significand_bits, sign as plus or minus one, biased_exponent, unbiased_exponent when defined, category, exponent_bias, source, message, and ordered steps. Failed validation—missing every source field, non-numeric decimal text, invalid hex digits, or non-binary characters—returns invalid_input and is never charged. Pricing is a flat $0.002 per successful request with unit item and zero marginal unit price.

Practical tips, limits, and related searches

Prefer precision sixty-four when your source is a JavaScript number or another binary64 host type, and switch to thirty-two only when you are matching float or float32 storage on the wire or in a file format. When round-tripping, encode a value then decode the returned hex with the same precision and confirm category and fields match. Keep patterns inside the published bit width; patterns longer than the selected precision are rejected rather than truncated, which would silently change the value. Remember that many decimals such as one tenth do not have a finite binary expansion; the hex you see is the correctly rounded IEEE encoding, not a bug. Use signed zero, infinity, and NaN examples when teaching specials: encoding zero and negative zero yields distinct sign bits and the categories pos_zero and neg_zero even though JSON prints both as zero and the tool exposes is_neg_zero for that case. This landing targets people searching for a float to IEEE-754 converter, single or double precision hex dump, or binary floating-point bit pattern tool; it is not a decimal-floating-point (decimal64) converter, not an arbitrary-precision calculator, and not a general hex-to-binary integer tool—use the sibling hex-to-binary capability when you only need integer base conversion. Failed tasks that return invalid_input do not consume prepaid balance, which keeps exploratory scripts cheap when learners deliberately send garbage strings.