{
  "id": "encoding/z-base-32",
  "family": "encoding",
  "slug": "z-base-32",
  "title": "z-base-32 (Zooko's human-oriented base-32)",
  "summary": "z-base-32 is Zooko O'Whielacronx's permuted Base32: the same 5-bits-per-character scheme as RFC 4648 Base32, but with the alphabet 'ybndrfg8ejkmcpqxot1uwisza345h769', chosen so the easiest-to-say and most frequent symbols map to the most common data, with no padding characters. It is used in Tahoe-LAFS, Mnet, and BitTorrent Mainline DHT contexts where a compact, transcribable, unpadded encoding is wanted.",
  "kind": "encoding",
  "aliases": [
    "zbase32",
    "z-base32",
    "Zooko base32",
    "human-oriented base-32"
  ],
  "status": "de-facto",
  "verification": "verified",
  "tier": "B",
  "source_url": "https://philzimmermann.com/docs/human-oriented-base-32-encoding.txt",
  "source_version": "z-base-32 specification (Zooko O'Whielacronx), retrieved 2026-05-29",
  "retrieved_date": "2026-05-29",
  "see_also": [
    "encoding/base32",
    "encoding/base32hex"
  ],
  "ext_type": "encoding@1",
  "ext": {
    "rfc": "z-base-32 — human-oriented base-32 (Zooko O'Whielacronx)",
    "charset": "US-ASCII",
    "algorithm": "z-base-32",
    "alphabet": "ybndrfg8ejkmcpqxot1uwisza345h769",
    "notes": [
      "Same 5-bit grouping as standard Base32, but a permuted alphabet 'ybndrfg8ejkmcpqxot1uwisza345h769' (index 0 = 'y', index 1 = 'b', ...). No '=' padding is ever emitted.",
      "The original spec is bit-oriented: it encodes an exact bit count, so a final partial 5-bit group takes only the bits present (e.g. 1 bit of 0 -> 'y', 1 bit of 1 -> 'o'; 2 bits '11' -> 'a'). Spec vectors include: 10 bits 1000000010 -> 'on', 20 bits 10001011100010001000 -> 'tqre', 24 bits 111100001011111111000111 -> '6n9hq', 24 bits 110101000111101000000100 -> '4t7ye'.",
      "This entry's executable test_vectors use the common BYTE-aligned profile (whole bytes in, trailing group zero-extended to 5 bits), which round-trips byte<->string; the bit-exact form above is documented for fidelity to the original spec.",
      "Alphabet ordering is NOT monotonic, so encoded strings do not sort in byte order (unlike base32hex). The point is human transcription, not sortability."
    ],
    "test_vectors": [
      {
        "input": "",
        "input_form": "utf8",
        "output": "",
        "output_form": "ascii",
        "algorithm": "z-base-32",
        "direction": "encode",
        "note": "Empty input -> empty output."
      },
      {
        "input": "f",
        "input_form": "utf8",
        "output": "ca",
        "output_form": "ascii",
        "algorithm": "z-base-32",
        "direction": "encode",
        "note": "Byte-aligned: 'f' (0x66) -> 'ca' (8 bits -> 2 chars, trailing 5-bit group zero-extended)."
      },
      {
        "input": "fo",
        "input_form": "utf8",
        "output": "c3zo",
        "output_form": "ascii",
        "algorithm": "z-base-32",
        "direction": "encode",
        "note": "Byte-aligned: 'fo' -> 'c3zo'."
      },
      {
        "input": "foo",
        "input_form": "utf8",
        "output": "c3zs6",
        "output_form": "ascii",
        "algorithm": "z-base-32",
        "direction": "encode",
        "note": "Byte-aligned: 'foo' -> 'c3zs6'."
      },
      {
        "input": "foobar",
        "input_form": "utf8",
        "output": "c3zs6aubqe",
        "output_form": "ascii",
        "algorithm": "z-base-32",
        "direction": "encode",
        "note": "Byte-aligned: 'foobar' -> 'c3zs6aubqe'."
      },
      {
        "input": "c3zs6aubqe",
        "input_form": "ascii",
        "output": "foobar",
        "output_form": "utf8",
        "algorithm": "z-base-32",
        "direction": "decode",
        "note": "Round-trip: z-base-32-decode(\"c3zs6aubqe\") = \"foobar\"."
      }
    ]
  },
  "updated": "2026-05-29T00:00:00Z"
}