Walendria Protocol
A minimal on-chain escrow table: the buyer funds, then releases to the seller or burns.
Formerly called Meridian Protocol / MP. New public name: Walendria Protocol / WP.
1. What is this?
Walendria Protocol is not a court. It is not a refund desk. It is not an auditor.
It is a table with one seller and one buyer/controller.
The buyer funds the table. The seller cannot withdraw early.
The buyer can release to the seller or burn to the dead wallet.
Release if satisfied. Burn if not.
2. Protocol vs Application — know the difference
WP is a base layer. It is a smart contract on Base that manages escrow tables. It does not have a dashboard, a web form, or an easy button. It is code.
Applications are built on top of the protocol to make it easier to use — they can add interfaces, wallets, token selectors, AI agents, or chat-based flows. Applications may have opinions about UX; the protocol does not.
The Protocol (WP)
- A smart contract on Base mainnet
- Stores escrow table state on-chain
- Enforces: fixed seller, fixed buyer, release or burn
- Does not care about UI, language, or user type
- Cannot be changed without a redeployment
- Source-verified and public
Applications (built on WP)
- A CLI tool that parses and formats commands
- A web dashboard for human buyers and sellers
- An AI agent that executes tables autonomously
- A chat bot that walks you through funding
- Any custom integration that calls the contract
- Can be built by anyone, upgraded independently
3. Who can use WP?
The original vision involves AI agents that transact autonomously — an agent can fund a table, check state, and release without a human clicking anything. That is powerful.
But the game theory works equally well for two humans making a deal. If you are a freelancer, a buyer, or anyone doing an internet transaction with a stranger, the escrow logic applies the same way.
The protocol does not distinguish between a human wallet and an AI agent wallet. It only asks: did the buyer/controller release or burn?
A buyer and seller agree on a deal in a chat. Buyer creates a WP table, funds it with USDC. Seller sees the funds are locked. Seller delivers. Buyer releases. Seller withdraws.
An AI agent is instructed to pay for a service. It creates a table, funds it, and releases after verifying the deliverable — without a human clicking anything at each step.
4. How to use it
Install the small WP live CLI. It can read the WP V3 contract and send Base transactions when you provide a signer.
Then use these shapes. First set a hot-wallet signer in your shell. The signer pays gas and signs actions. WP_YES=1 is for AI agents/scripts that cannot type the confirmation prompt. Do not use a valuable wallet key.
+26:USDC means "fund 26 USDC." -0.05:ETH means "withdraw 0.05 ETH." Release and burn always affect all assets in the table at once.
wp seller buyerCreates a table. Uses WP_PRIVATE_KEY from your shell for gas/signing.
wp contract:tableidShows table state: seller, buyer/controller, balances per asset, state, withdrawable amounts.
wp table +amount:TOKENAdds funds from the WP_PRIVATE_KEY signer. TOKEN is ETH, USDC, or a contract address.
wp table releaseBuyer/controller signer releases all assets to seller-withdrawable.
wp table burnBuyer/controller signer sends all assets to dead wallet. Destructive settlement, not a refund.
wp table -amount:TOKENSeller signer pulls a specific asset after release. Specify which token to withdraw.
5. Why the game theory works
WP reduces trust friction by removing profitable shortcuts. Seller cannot steal early. Buyer cannot get work and refund himself. The buyer's only final buttons are release or burn.
The buyer cannot profit from cheating.
Traditional refunds feel fair, but they also arm dishonest buyers. A buyer can receive work, complain, and claw money back. WP removes that profit path. If the buyer refuses to pay, the buyer must burn too.
The seller cannot self-release.
The seller is receiver, not judge. A receiver should not approve their own payment. The seller waits, then withdraws only after release.
Random money should not control the table.
The buyer/controller is fixed at creation. A seller or stranger should not be able to dust a table and steal control. Fixed authority prevents "largest funder wins" and "first funder wins" attacks.
Timelock is mandatory, immutable, and visible.
Old WP doctrine avoided hidden automatic settlement. This redeploy keeps that caution, but adds explicit mandatory immutable timelock: every table is created with a public future timestamp, and anyone can claim auto-release after it expires. The timestamp cannot be moved or disabled after creation.
The protocol stays small on purpose.
- Buyer funds, because the buyer is the source of the committed risk.
- Buyer releases or burns, because the buyer is the controller.
- Seller withdraws after release, because the seller is the receiver.
- Nobody refunds by default, because refunds create profitable fraud.
- Nobody judges by default, because judgment creates attack surface.
6. V3 live doctrine
Current WP V3 tables are asset-agnostic at creation. A buyer can fund the same table with native Base ETH and/or ERC20 tokens. Each asset is tracked separately.
marker: 0x000...000Native asset marker inside the accounting model. Use +amount:ETH in CLI grammar.
fundToken(tableId, token, amount)Buyer approves the WP contract first, then funds in raw token units. Use +amount:TOKEN.
1 USDC = 1 (use "26:USDC")Applications handle decimal conversion. The protocol stores raw units. A good app shows human-readable amounts.
current fee: 0%; owner-capped max: 1%Current release fee is 0 bps. The owner can update it, but the contract hard-caps protocol fee at 100 bps / 1%.
7. Live contract references
Always verify chain state before using real money. These are public references, not a promise that every wrapper/CLI has caught up.
8. What WP is not
WP is not an auditor.
WP is not a judge.
WP is not a refund service.
WP is not legal protection.
WP is not scam-proof.
WP is not a guarantee of delivery.
WP is not a guarantee of fairness.
WP is not a hidden buyer clawback.
WP asks one hard question: did the buyer/controller release or burn?
9. Operations
10. Begin with one clean table.
One seller. One buyer. One public vault. One final choice.
Release if satisfied. Burn if not.
Create the first table