ipfs cid in base58 qm
An IPFS CID is not a raw hash: it is a multihash (algorithm code plus digest length plus the digest itself) encoded as text. CIDv0 uses Base58btc, the same 58-symbol alphabet as Bitcoin addresses, which is why it almost always adds up to 46 characters starting with "Qm". CIDv1 swaps that encoding for Base32, and the reason is not cosmetic.
The 34 bytes (2 prefix bytes plus 32 digest bytes) turn into 46 Base58btc characters; the two fixed leading bytes, 0x12 and 0x20, are what always produces "Qm" at the start.
Decoding confirms the structure: the first two bytes come back as 12 20 (SHA-256 code, length 32), and the following 32 bytes are the raw digest, with no data about version or codec at all.
The same 32 digest bytes produce a completely different-looking CID because the version, the codec and the alphabet change, not because the referenced content changed.
Base58 drops 0 (zero), O (uppercase o), l (lowercase L) and I (uppercase i) because they look alike in common fonts and are hard to tell apart when read aloud. That reduces transcription errors, which is critical for wallet addresses, where one swapped character can mean permanently lost funds.
Because a CIDv0 is a SHA-256 multihash (2 fixed header bytes, 0x12 and 0x20, plus a 32-byte digest) encoded in Base58btc, and those two header bytes always produce the characters "Qm" as the opening, no matter the file. Changing the hash algorithm would change the prefix; changing only the file content would not.
No. Both can carry the exact same SHA-256 digest, so they reference the same content. What changes is the representation: CIDv0 is always Base58btc with no version or codec fields, CIDv1 adds those fields and uses Base32 by default. Modern IPFS clients accept and resolve both formats to the same content.
Because DNS domain names are case-insensitive and normalise everything to lowercase before resolving. A CIDv0 in Base58btc deliberately mixes upper and lowercase letters (that is part of the 58-symbol alphabet), so two different CIDs could collide once DNS folded the case. CIDv1's lowercase Base32 removes that ambiguity.
123456789ABCDEFGHJKLMNPQRSTUVWXYZabcdefghijkmnopqrstuvwxyzEverything is computed in your browser, including the SHA-256 behind the checksum. No key or address leaves this device.