The latest articles on DEV Community by levithefirst (@levithefirst). https://dev.to/levithefirst https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4232395%2Fff1c53e0-21f2-4d42-aef4-d80699a3e1a5.jpg https://dev.to/levithefirst en levithefirst Sat, 07 Nov 2026 16:05:41 +0000 https://dev.to/levithefirst/clew-remembering-the-retry-that-worked-9gj https://dev.to/levithefirst/clew-remembering-the-retry-that-worked-9gj <p><em>Tatum × Walrus Memory “Total Recall” hackathon. Live at <a href="https://clew-eight.vercel.app" rel="noopener noreferrer">clew-eight.vercel.app</a>.</em></p> <p>A transaction reverts. The wallet says “execution reverted”, maybe with a reason string. Someone tweaks a number, resubmits, and it goes through. A week later the same wallet hits the same contract with the same error and starts from zero, because nothing kept the fix.</p> <p>If you build agents that send transactions, that loop burns retries and leaves you guessing which change actually landed. Wallet users hit the same wall: they pay gas twice and still walk away without knowing whether a slippage limit or a deadline was the thing that moved. Hackathon judges can check a concrete claim on real Mainnet data: a stored memory changes what you do on the next failure.</p> <p>Clew watches that loop and stores it.</p> <h2> What Clew does </h2> <ol> <li>A transaction fails. Clew reads the receipt and decodes why.</li> <li>The same wallet retries the same contract and succeeds. Clew diffs the two attempts and writes what changed to Walrus Memory on Mainnet.</li> <li>The next time that wallet fails on that contract with a similar error, Clew recalls the memory and returns the prior successful transaction plus a concrete next action.</li> </ol> <p>With an empty memory it can only say the transaction reverted. Once a memory exists, it can name which words in the call changed, and point at the retry that landed.</p> <p>The public website never writes. Both the Try it box and the agent endpoint only read.</p> <p>If you run an agent that retries failed swaps, send the failed hash to <code>/v1/recall</code> and you get the last path that worked for that wallet and contract.</p> <h2> How Tatum is used </h2> <p>Every chain read goes through Tatum, so a pasted failure is decoded from a live receipt.</p> <p>On Base, Ethereum and Polygon that means the JSON-RPC gateways:</p> <ul> <li> <code>eth_getTransactionReceipt</code> for status. Clew treats <code>0x0</code> as a failure and only accepts a later <code>0x1</code> transaction as the retry.</li> <li> <code>eth_getTransactionByHash</code> for calldata, gas limit, gas price, value and nonce. <code>eth_getBlockByNumber</code> for the timestamp.</li> <li>Revert decoding via <code>debug_traceTransaction</code> with <code>callTracer</code>, falling back to an <code>eth_call</code> replay at <code>block - 1</code>. Clew decodes <code>Error(string)</code> (<code>0x08c379a0</code>), <code>Panic(uint256)</code> (<code>0x4e487b71</code>) and custom error selectors, then normalizes them into a signature such as <code>revert:too_little_received</code>.</li> <li> <code>eth_getBlockReceipts</code> to find the successful retry from the same wallet to the same contract after the failure.</li> </ul> <p>The free plan allows 3 requests a second. Clew spaces its calls about 360 ms apart and backs off on HTTP 429.</p> <p>Sui uses a different door on the same provider. Tatum’s Sui gateway serves GraphQL only (<code>POST https://sui-mainnet.gateway.tatum.io/graphql</code>). A first attempt at <code>sui_*</code> JSON-RPC came back HTTP 404, so the adapter was rewritten against GraphQL. One transaction query returns the sender, gas budget and price, status, checkpoint, Move calls, Pure inputs, and <code>executionError</code> (the module, function and abort code where it died). That becomes a signature like <code>moveabort:errors::err_amount_out_slippage_check_failed:1</code>. Finding the retry is supposed to take one filtered query: same sender, same <code>package::module</code>, after the failure’s checkpoint.</p> <p>That GraphQL path was verified on Sui Mainnet on 31 October 2026: the gateway answered at checkpoint 340173364, recent programmable transactions decoded, and two real fail → success pairs were written to Walrus. On 7 November 2026, single-transaction reads still worked and the live <code>/v1/recall</code> endpoint answered two digests with the same structured shape as Base. The list query (<code>transactions</code>) returned “Failed to list transactions” that day, so a <em>new</em> pair could not be discovered then.</p> <h2> How Walrus Memory is used </h2> <p>Memories go to the Walrus Memory Mainnet relayer (<code>https://relayer.memory.walrus.xyz</code>) in namespace <code>clew</code>, through <code>@mysten-incubation/memwal</code>. Later failures from the same wallet can pick up the retry that already worked because that write lives there.</p> <p>A write is <code>remember(text, { idempotencyKey: "chain:failedTx:successTx" })</code>, then <code>waitForRememberJob</code> until a <code>blob_id</code> comes back. If that confirmation never arrives, the pair is not treated as stored. A read is <code>recall</code> with a query built from the new failure’s chain, contract, wallet, error and selector. Recall is semantic, so Clew then applies a strict filter: exact chain, exact contract, exact wallet, and a similar error. Another wallet’s path is never reused.</p> <p>The memory text is plain and factual, tags first so both recall and the filter have something to hold on to:</p> agents hackathon software web3