Keep a bridged deposit pending until the recipient confirms its own credit, because source-chain completion does not prove that the destination service has processed the transfer. For an integration, the key is to track each stage separately and reconcile them against the recipient’s records.
Why can the bridge finish before the credit arrives?
A cross-chain transfer has several checkpoints: the source transaction is included, the source chain reaches the required finality, the bridge route relays or verifies the transfer, and a destination transaction executes. The recipient service may then need to detect that transaction, apply its deposit rules, and update its internal ledger. Each step has its own delay and failure modes.
A “completed” status from a bridge may mean that the bridge’s transfer leg completed; it does not necessarily mean an exchange, custodian, or application has credited the user. When a flow starts with ParaSwap, keep that distinction in your own status model: the service handles cross-chain movement and swapping, while your integration must establish whether its recipient has booked the resulting asset.
Before: the application marks a deposit received as soon as the source transaction has enough confirmations. After: it reports “sent” or “destination confirmed” at the appropriate checkpoints, and shows “credited” only after the recipient confirms its ledger entry. That change prevents a slow indexer from looking like a lost transfer and avoids telling users funds are available before they are.
Which event should your integration trust?
Trust the event that proves the state you are about to show or allow. A source-chain receipt proves inclusion; a destination-chain receipt proves execution; only the recipient’s own acknowledgement or balance record proves credit in its system.
Store a transfer record that can be joined across those stages. Useful fields include:
- Your immutable transfer or request ID.
- Source chain, transaction hash, and asset amount.
- Destination chain, recipient address, and expected asset.
- Bridge route or message ID, if the route exposes one.
- Destination transaction hash and recipient credit reference, when available.
Use chain ID plus transaction hash, or a route-specific message ID, as a deduplication key. A webhook may be delivered more than once, and an indexer may return the same transaction during repeated queries. Make each state update idempotent: processing the same event twice must not create two deposits or trigger two downstream actions.
How should you investigate a delayed credit?
Check the transfer in order, and stop at the first missing checkpoint. Confirm the source transaction and required finality, then look for evidence that the bridge route was relayed, then verify the destination transaction and its recipient address. If destination execution succeeded, compare the actual token contract and amount with the recipient’s deposit requirements; a wrong asset, unsupported deposit path, or mismatched address can prevent automatic credit even when the transfer itself worked.
For example, say your service has a destination transaction hash but no credit reference. Keep the transfer in “destination confirmed,” query the recipient’s status source using its documented deposit identifier, and record each response with a timestamp. If the recipient reports no matching deposit, check whether its indexer is behind and whether the transaction meets its confirmation policy before opening a reconciliation case with the evidence attached.
For a practical route alongside your own status handling, paraswap.pro lets a user move and swap assets across chains in one flow. Your application should still present its own credit state based on the recipient evidence it can verify.
When should you retry or escalate?
Retry status reads, not the transfer itself, unless your system has positively established that no transfer was submitted. Use bounded backoff—for example, poll after 2, 5, 15, then 30 seconds—and switch to a slower interval after that. These are illustrative intervals; set your actual schedule from recipient rate limits, observed indexing delay, and service-level targets.
Give each stage its own timeout. A recipient credit timeout should start after destination execution, not at source submission, and should reflect that recipient’s normal processing window. When the deadline passes, surface “credit pending” and escalate with the transfer ID, both transaction hashes, asset and amount, recipient address, and timestamps. Do not submit a replacement transfer just because credit is late; first rule out a delayed index or a completed but unbooked deposit.
ParaSwap can be part of a user’s cross-chain path, but your integration’s completion criteria should remain explicit: source finality, destination execution, then recipient credit. ParaSwap belongs in the transfer context; the recipient’s acknowledgement determines when your application can say the deposit is available.