Our Jobs API is available in the EVM, UTXO, and SVM (Solana) namespaces.
Purpose
Jobs are designed to handle backfills of large accounts, pulling a set of transactions within a defined time range as fast as possible, or any other scenario where a large amount of throughput is required and the regular txs endpoint aren't fast enough.
They also expose additional features and formatting options that aren't available in the regular txs endpoints.
For example, you can request transactions to include token prices for every transfer, or you can request a reconciliation to be performed (this will check the onchain balances for every token against the accumulated balance calculated from our API output).
We encourage you to favor jobs for most production-level integrations. They will be faster, more performant, and will offer more features.
How it Works
Internally, jobs are streams managed entirely by us, with all the infra-level intricacies abstracted away from the user.
To you, they're simply HTTP endpoints where you can request data for a particular account and a given time range, and once data is ready it will be available when you call again.
You can also pass a webhook URL to get notified once the data is available.
On Solana, in addition to fetching all the data for your requested account, we also perform a recursive search to find all associated token accounts (both active and closed), and also return their transactions as part of the job. This saves you from having to code your own logic for fetching a fully-complete Solana history (see associated article here).
If you start a job by mistake and want to stop it before it processes all transactions (and deducts compute units), there's a DELETE endpoint available on every namespace where you can pass the job ID you want to cancel.
Billing
Once you start a job, it will run through completion and you will be charged CUs based on the amount of transactions you process. The CU charge per transaction is the exact same as if you were calling the regular tx endpoint for that particular namespace.
You're charged once per transaction processed. The job's data is then stored and available to be fetched as many times as you need, for a maximum of 2 weeks.