SlotScribe is a verifiable execution recorder:
- Your agent runs normally (build/sign/send). SlotScribe does not custody keys.
- It records an off-chain trace (
intent,plan,toolCalls,txSummary). - It computes
payloadHash = sha256(canonicalize(trace.payload)). - It anchors that hash on-chain via Memo:
SS1 payload=<hash>. - Anyone can independently verify by signature/hash.
As agents execute real-value actions, trust must be verifiable:
- What exactly happened?
- Can the agent rewrite the story later?
- How can users or other agents audit behavior?
SlotScribe turns claims into cryptographic receipts.
- Docs page:
/docs - Verify page:
/verify - Public Trace API:
https://slotscribe.xyz/api/trace - AI Prompt guide:
./docs/AI_Agent_System_Prompt.md
npm i slotscribe@latest @solana/web3.js
# or
pnpm add slotscribe@latest @solana/web3.jsCheck latest version:
npm view slotscribe versionimport { SlotScribeRecorder, getConnection } from 'slotscribe';
import { Keypair, SystemProgram, Transaction } from '@solana/web3.js';
const cluster = 'devnet';
const connection = getConnection(cluster);
const payer = Keypair.generate();
const recorder = new SlotScribeRecorder({
intent: 'Quickstart transfer test',
cluster,
});
recorder.setTransferTx({
feePayer: payer.publicKey.toBase58(),
to: payer.publicKey.toBase58(),
lamports: 1000,
});
const tx = new Transaction().add(
SystemProgram.transfer({
fromPubkey: payer.publicKey,
toPubkey: payer.publicKey,
lamports: 1000,
})
);
const signature = await recorder.sendTransaction(connection, tx, [payer], {
autoUpload: true,
baseUrl: 'https://slotscribe.xyz',
});
const trace = recorder.buildTrace();
console.log('signature=', signature);
console.log('payloadHash=', trace.payloadHash);
console.log('verifyUrl=', `https://slotscribe.xyz/verify?cluster=${cluster}&sig=${signature}&hash=${trace.payloadHash}`);pnpm install
pnpm devGET /api/verify?cluster=...&signature=...&hash=...POST /api/traceGET /api/trace/{hash}POST /api/trace/{hash}
- SlotScribe verifies integrity (hash consistency), not business truthfulness of off-chain tools.
- Use policy gates (allowlists/limits/slippage checks) for production safety.
src/slotscribe/: SDK coreapp/: Viewer UI + API routeslib/: verifier/storage helpersscripts/: demos and validation scriptsdocs/: guides and prompts
MIT