# Status and costs reference

## Status rule

Separate these states:

- `live`: the exact deployment identity exists on the named chain and a current read proves the interface;
- `prepared`: source and call instructions exist, but required deployment inputs or deployment are incomplete;
- `unknown`: a current read was not possible;
- `stale`: only an older receipt or prior run is available.

Never promote `prepared`, `unknown`, or `stale` to `live`.

## Public-good rule

Uncle Matt funds deployment and intends the deployed interfaces to remain permissionless. The
Randomifiers charge no subscription, account, API-key, provider, protocol, or license fee.

Current Solana wording: “Live on Solana mainnet. Free to call; normal network fees apply.”
For Robinhood Chain, retain its separate in-development status.

This does not erase third-party costs:

- network transaction fees belong to the chain;
- RPC plans and rate limits belong to the selected RPC provider;
- integrator hosting and application costs belong to the integrator.

For Hood, direct off-chain `eth_call` does not submit a transaction. Contract-to-contract calls consume
gas inside their parent transaction. Solana transaction-level calls require the ordinary network fee.

## Dated Solana measurements — 2026-10-14

Local SBF execution with synthetic feed accounts measured 14,430 CU direct and 21,628 CU through a validating CPI consumer. Local standalone fee quotations were 5,000 lamports with no priority price. These are fixture measurements, not worst-case or guaranteed production prices. Measure the whole consuming transaction; CPI uses that transaction budget. Devnet deployment proof and public-mainnet feed observations are separate evidence lanes.

## Claim rule

Use “permissionless public-good formula oracle” only with a current deployment status beside it.
Use “no Randomifier service fee,” not the broader and usually false claim “no cost of any kind.”

## Reviewed release rehearsal — 2026-10-15

The 42,184-byte SBPFv2 release reproduced all 95 result bytes against seven real mainnet feed account snapshots in a local validator at 14,138 CU. The snapshots were sequential reads. The retained sanitized proof is https://bttlabs.fun/randomifiers/evidence/LOCAL_MAINNET_CLONE_PROOF_2026-10-15.public.json. Keep these release measurements separate from the October 14 devnet artifact and synthetic fixture measurements.