Credits
An estimate for relative usage, weighted from the one number Bright Data actually publishes.
Constraint
Bright Data publishes no per-operation credit price list. Every figure on this page, in the CLI, and in the cockpit is an estimate for relative usage — useful for "which collector is expensive to keep healthy", not usable as a bill.
Where the weights come from
The one fact that is actually documented: a single-page run costs one credit regardless of row
count. Everything else is an AI-Flow job with no published cost at all, so Molt turns that one fact
into a fixed ratio:
| Command kind | Weight |
|---|---|
run | 1 |
approve | 1 |
reject | 0 |
heal | 12 |
create | 12 |
unknown | 0 |
heal and create are weighted twelve times a run because they are the AI-Flow jobs — slow,
capacity-limited, and (by every account of how these systems are usually priced) the expensive
operations in the fleet. The weights are overridable in code the moment a real number is published, or
the moment your own invoice tells you what they actually are.
Reading it
molt creditsmolt credits chaosFleet-wide with no argument, or scoped to one collector. Classification works from a command's own
argv, never from a side channel — so the ledger is reconstructable from the commands table alone,
with no separate bookkeeping to drift out of sync with it.
Every place a credit figure is shown — the CLI, the Fleet page, an incident's cost-of-silence line — carries the same tooltip: estimated relative usage, not a price.