> ## Documentation Index
> Fetch the complete documentation index at: https://docs.manthan.systems/llms.txt
> Use this file to discover all available pages before exploring further.

# Replay

> Deterministic reconstruction of a recorded decision

**Replay answers: can this outcome be reproduced from what was recorded?**

Replay deterministically re-evaluates a Business Transaction's recorded policy version against its recorded signals, and confirms whether the outcome matches what's in the Execution Trust Record. This is only possible because Policy and Decision are both deterministic and immutable — see [Concepts → Policy](/docs/concepts/policy) and [Concepts → Decision](/docs/concepts/decision).

## Result model

```ts theme={null}
interface ReplayResult {
  businessTransactionId: string;
  trustRecordHash: string;
  verified: boolean;
}
```

## Triggering a replay

<CodeGroup>
  ```ts TypeScript SDK theme={null}
  const replay = await client.replay(businessTransactionId);
  ```

  ```python Python SDK theme={null}
  replay = client.replay.replay(business_transaction_id)
  ```

  ```bash REST theme={null}
  curl -X POST http://localhost:3000/replay \
    -H "Content-Type: application/json" \
    -d '{ "businessTransactionId": "txn-001" }'
  ```
</CodeGroup>

## Why this matters

Replay is what lets a third party — an auditor, a regulator, a new engineer on the team — independently confirm that a past decision was correct, without having to trust that the original runtime behaved honestly. It's the practical payoff of the "deterministic execution" and "immutable trust artifacts" principles in [Architecture → Overview](/docs/architecture/overview).

Replay logic lives in the [`replay`](/docs/packages/replay) package.