Glossary · The money
Cost attribution
Tracing every dollar a task spent back to the stage and agent involved, and the model that ran it.
What it is
Cost attribution means every unit of model spend is tied to the exact session that caused it, and a session that belongs to a task rolls up to that task. The result is a real number per task instead of an estimate spread across one monthly bill.
Why it matters to you
Without attribution, an AI bill is one big total with no way to tell which kind of work is actually expensive. With it, you can see that planning costs more than coding on a typical task, or that one repeatedly reworked task is quietly driving the spend, and act on either fact directly.
How it works in CodeHerder
CodeHerder records every agent turn and attributes it to the session that ran it, and that session rolls up to the task it belongs to. `ch task show` prints a task's cost inline, broken down by stage. `ch costs` and the web app's Cost breakdown page split the same spend by agent, by model, or over any time window.
What people get wrong
People sometimes assume cost tracking means CodeHerder is billing them for tokens. You bring your own model provider key, and token spend goes straight to that account. Cost attribution only makes visible the spend that's already happening; it adds no separate charge on top of it.
Keep going
Related terms
Rework
A task sent back to the build stage because review or verify found something wrong, at extra cost.
Concurrent agents
How many agent sessions can be running for your workspace at the same time, the actual unit CodeHerder's plans are metered on.
Context window
How much text a model can actually hold in mind at once, and one of the biggest levers on what a task costs to run.
Round up your herd.
Bring every human and every agent onto one table. Watch the work move. Costs update as it happens.