StarkNet Namespace - Interoperable Address

Namespace Reference

ChainType binary key: 0x0003

CAIP-104 namespace: starknet

Chain reference

See this namespace’s CAIP-2 profile. Chains are identified by string identifiers (e.g. SN_MAIN, SN_GOERLI) that are encoded as field elements on the chain but expressed as ASCII strings in CAIP-2.

Text representation

<chain_id_string>

Where <chain_id_string> is the case-sensitive chain identifier (e.g. SN_MAIN, SN_GOERLI).

Note: Per CAIP-350, the full chain identifier is starknet:<chain_id_string> (e.g., starknet:SN_MAIN, starknet:SN_GOERLI).

Text representation -> customary (CAIP-2) conversion

The text representation (chain reference) is the same as the chain reference in the CAIP-2 chain identifier; no conversion is needed.

Customary (CAIP-2) conversion - text representation conversion

The chain reference in the CAIP-2 chain identifier is the same as the text representation; no conversion is needed.

Binary representation

The chain reference is the UTF-8 encoding of the chain ID string. Length is variable (e.g. 7 bytes for SN_MAIN, 9 bytes for SN_GOERLI). The ChainReference field in binary Interoperable Addresses must carry this byte sequence; higher-level framing (e.g. length prefix) is defined by ERC-7930 / CAIP-350.

Text -> binary conversion

Encode the chain ID string as UTF-8 bytes.

Binary -> text conversion

Decode the bytes to the chain ID string (for all current identifiers, the bytes are UTF-8/ASCII).

Examples

Addresses

See this namespace’s CAIP-10 profile. StarkNet addresses are 32-byte field elements, represented as 0x followed by 64 hexadecimal characters (zero-padded), with optional EIP-55 checksum.

Text representation

<address>

Where <address> is the 32-byte field element as in CAIP-10: 0x followed by 64 hex characters (EIP-55 checksum recommended), e.g. 0x02DdfB499765c064eaC5039E3841AA5f382E73B598097a40073BD8B48170Ab57.

Text representation -> native representation conversion

Strip the 0x prefix and decode the 64 hex characters to 32 bytes. Checksum validation (if present) follows EIP-55.

Native representation -> text representation conversion

Encode the 32 bytes as 64 hex characters with 0x prefix. Apply EIP-55 checksum for the canonical text form.

Binary representation

The address is stored as the raw 32 bytes (big-endian), as in the native field element representation. No length prefix is needed for fixed-size addresses.

Text -> binary conversion

Remove the 0x prefix and decode the 64 hex characters to 32 bytes (via RFC-4616 base16).

Binary -> text conversion

Encode the 32 bytes as 64 hex characters and prepend 0x. Apply EIP-55 checksum casing for the canonical form.

Examples

Error handling

When converting from CAIP-2 to this profile, the chain reference is fully determined by the chain ID string, so no loss of information occurs. If a chain ID is unknown or the UTF-8 encoding is invalid, implementations should signal an error.

Implementation considerations

The interoperable format uses the chain ID string for text representation and its UTF-8 encoding for binary. Chain IDs may be resolved via the starknet_chainId method when needed (e.g. to discover the identifier for a connected node). Addresses are always 32 bytes; leading zero in hex is normal for field elements below 2^255.

References

Citation

Please cite this document as:

Orca, Mono, "namespaces/starknet-caip350: StarkNet Namespace - Interoperable Address [DRAFT]," Chain Agnostic Namespaces, starknet-caip350, January 2026 / . [Online serial]. Available: https://github.com/ChainAgnostic/namespaces/starknet-caip350.md