Ship token launches without
leaking your cap table.

Confidential vesting, airdrops, and disperse on FHEVM. Three product stories. One protagonist, Aria. End-to-end on Sepolia in your browser.

Same state, three lenses
Public · Etherscanamount = ???
0x9f44a9e5b7b8c0d1e2f3a4b5c6d7e8f90a1b2c3d4e5f6a7b8c9d0e1f2a3b4c5d
Aria · operator ACLdecrypted
1,000,000 ARI total
Allocated to Mira · vested + claimable
Mira · recipient ACLdecrypted
262,328 ARI vested
6-month cliff cleared · linear unlock

Same ciphertext, three decrypted views. That's the FHE pitch made visceral.

Why @tokenops/sdk

The smallest surface area that respects FHE.

Encapsulates Zama's FHEVM stack, ACL grants, simulate-vs-receipt handles, KMS proofs, scaled integers, behind a viem-grade typed API.

Confidentiality at the math layer

Amounts are euint64 or euint128 ciphertexts depending on product — airdrop and disperse use euint64; vesting allocation, settled, and reserved amounts use euint128. Per-recipient amounts (airdrop/disperse handles, vesting allocation and settled) carry per-recipient ACL grants; vesting reserve counters stay contract-internal, while the disperse token-fee reserve is decryptable by an authorized fee-collector or admin via accessEncryptedFeeReserve. No off-chain spreadsheet, no privileged custodian, no trust-me API.

Same state, three lenses

Operators audit their roster. Recipients see only their own slice. Public sees a handle. One on-chain truth, three views.

Viem-first, typed end-to-end

8 branded primitives keep tx hashes, signatures, and encrypted handles from silently swapping. Pattern Z, baked in.

Sepolia today, mainnet next

fhe-disperse is live on Sepolia + Ethereum mainnet. fhe-vesting + fhe-airdrop are audited by OpenZeppelin and ship to mainnet once deployment and Zama KMS mainnet readiness land. Drop in alongside your wagmi stack.

Building against the SDK?

The Zama bounty cohort opens against this published surface. Skim the stories for product intuition, jump to Quickstart for the install path, or read the Concepts for the mental models.