This is your own accounting — what your batched jobs actually cost, and what the same jobs would have cost run synchronously. It is personal to your key, so there is no anonymous version. Send your jobs to POST /v1/calls/complete and this ledger fills in, month by month, with money you have actually saved.
Retrospective on imported history
The retrospective replays the real should-batch decision over the jobs you ran before you started batching, and shows what it would have saved. It reads your own imported history, so it needs your key.
Why you can trust these numbers
Every figure here is your own — measured on the jobs you actually ran, not a marketing estimate. The realised ledger prices each batched job the same way the rest of batchwatch does, and the retrospective replays the exact should-batch decision the live API returns, so this page can never claim a saving the API would not. A saving we cannot price is shown as absent, never as zero. That is the whole point of a ledger: it is a record you can take to a meeting.
Read the /v1/savings and /v1/retrospective reference · How we measure