FAQ

Common questions from buyers, builders, and reviewers. Pointers go to the deeper docs.

Why not an L3?

Short answer: every privacy L3 we have studied — including the canonical Aztec Connect post-mortem — shows the same failure surface (centralized sequencer, bridge UX, new-chain cold start, single-team maintenance burden). zBase is an application on Base so it inherits Base's USDC liquidity, wallets, and CEX integrations rather than asking users to bridge into a fresh privacy chain.

See Why a facilitator on Base for the full failure-mode analysis.

Why not an L3 like shh.gg or Aztec?

Same answer as "Why not an L3?" above — the canonical failure modes apply identically whether the L3 is shh.gg, Aztec, or a hypothetical new privacy rollup. zBase's choice to live on Base is a deliberate avoidance of sequencer centralization, bridge UX tax, and cold-start anonymity sets. See Why a facilitator on Base for the full analysis.

How do I integrate stealth recipients as a provider?

The SDK + registry walkthrough is in Stealth addresses.

What env vars do I need?

The canonical list is .env.local.example at the repo root. A few callouts:

• ZBASE_MIN_DEPOSIT_DELAY_SECONDS is the correct env var name for the minimum-deposit-delay window used by both the facilitator and the decoy scheduler. The codebase historically used MIN_DELAY_SINCE_DEPOSIT in prose (you'll see this in older runbooks) — treat that as a legacy alias for the same concept. The actual env var the code reads is ZBASE_MIN_DEPOSIT_DELAY_SECONDS.
• POSTMAN_PRIVATE_KEY, HYPERSYNC_TOKEN, RPC URL, and MAX_DAILY_BUDGET_USD are the other must-set values for a Sepolia push.

Why not just KYC at deposit?

Pre-deposit KYC drops the anonymity set to "users who passed our specific KYC vendor" and links a real identity to the deposit address. The Privacy Pools design (Buterin, Sun, Hu, Illum, Mahmoody 2023) trades a one-time deposit-time label check for permanent payment-time anonymity without asking the depositor to identify themselves. See Why a facilitator on Base.

Is this Tornado Cash with extra steps?

No. Tornado Cash had no ASP — sanctioned and clean funds were indistinguishable at withdraw time, which is why it was sanctioned in August 2022. zBase requires every withdrawal proof to show membership in both the state tree (any deposit) and the ASP root (approved deposits). Sanctioned wallets can still deposit but cannot privately withdraw. See Trust model §3.

Why both ASP and stealth addresses?

ASP defeats the prosecutor ("did you handle sanctioned funds?"). Stealth defeats the attacker ("which provider is wallet X paying?"). They sit at different lifecycle points by necessity — ASP at deposit time, stealth at payment time — and you cannot merge them without destroying privacy. See The two gates.

Why is the postman a single key today?

It is a known centralization. The 3-of-5 ThresholdEntrypoint is written and tested, but the contract bakes signer addresses as immutable constants — so external signers must be committed before it can deploy. This is on the Roadmap.

Is mainnet live?

Yes. Private x402 payments work end-to-end on Base mainnet today — deposit gaslessly, settle from the pool via a ZK withdrawal, and the seller delivers (200). Proven against real Coinbase-CDP sellers: BlockRun ($0.028), Nansen ($0.01), and MetaLend ($0.001).

It's an open pilot: the anonymity set is still filling toward its 30-depositor privacy floor, so payments settle and deliver but aren't yet crowd-anonymous (result.privacy.private tells you on each payment). The 30-depositor floor is necessary but not sufficient — a unique amount or timing can still narrow the set (see the Threat model). Remaining hardening — decentralizing the postman to a 3-of-5 threshold and an external audit before scaling TVL — is roadmap, not a blocker to paying today.

Does zBase claim complete anonymity?

No. The claim, proven on Base mainnet, is that the provider-payment transaction no longer names the payer wallet. Network metadata without Tor, behavioral fingerprinting, provider-side surveillance, and small anonymity sets all remain open. The honest disclosure surface lives in Trust model and Threat model.

How big is the anonymity set?

Dynamic. The 2026-07-11 Base Sepolia reference run is the current timing proof; query /api/anonymity-set for the live count. The set grows with deposits and does not shrink when funds are reclaimed (commitments stay in the Merkle tree forever, by design).

What protects me against zBase itself?

Threat	Mitigation
zBase rewrites the recipient	Recipient is bound into the proof's context public signal. Any rewrite breaks verification.
zBase logs your IP	The metadata-hygiene middleware strips identifying headers before any handler sees them.
zBase forges proofs	Cannot — the Groth16 verifier is on chain. Without the trusted-setup toxic waste, a forged proof is computationally infeasible.
zBase censors your deposit	Today: possible (postman is single key). After the threshold deploys: requires 3-of-5 collusion.

Where do I see what just shipped?

See CHANGELOG.md — the canonical release history + current state (fees).

Where is the API surface?

Wallet backend API is the modern entry; the legacy API reference lists every endpoint with full payloads.