> ## Documentation Index
> Fetch the complete documentation index at: https://docs.staffer.com/llms.txt
> Use this file to discover all available pages before exploring further.

# How pricing works

> Prepaid credits, one shared pool per organization, debited per AI action as it happens.

Staffer runs on prepaid credits, not seats. Your organization buys credits (or starts with \$100 free), and each AI action debits the pool the moment it runs.

## One pool for the whole organization

Credits sit in a single pool per organization, not per workspace. Every workspace your organization has draws from the same pool. An organization admin can set a monthly spend cap on an individual workspace to stop one team from using up the shared pool, but there's no separate credit balance per workspace. See [Credits, top-ups and spend caps](/admin/credits).

## What uses credits

Credits pay for AI and vendor work, grouped here by what you'd recognize in the app:

| Group | What's included |
| - | - |
| The agent and chat | Chatting with the agent on a job, and the **Ask anything** page chat on Jobs, Talent pool, and Applications |
| Searches and scoring | Running a candidate search, and scoring or rescoring candidates against a brief |
| Listings and drafting | Generating or translating a job listing, drafting or rewriting a message, naming a search, drafting a message template or career page copy |
| AI interviews and interview notetaking | The AI interviewer's conversation and voice, and recording and transcribing a human interview |
| Contact lookups | **Request contact info** on a candidate's profile |
| Research | Researching a company for workspace knowledge, and the web research the agent runs while it searches |
| Policy checks | The compliance check that runs before certain AI actions, when **Live policy enforcement** is on |

Every one of these is metered by the work it actually takes. A short search costs less than a long one, and a five-minute AI interview costs less than a twenty-minute one. See [What uses credits](/pricing/what-you-pay) for how each is charged.

Credits are checked before an action starts, not while it runs. A search or generation already in progress finishes even if it drains your organization's credits to zero along the way. The next action after that is what gets refused.

## What's free

Candidate-facing chat doesn't touch your credits at all. A candidate chatting on the apply page, in the candidate portal, or through a share link is paid by Staffer, not your organization, and it keeps working even if your organization is out of credits. AI interviews are different: they use your credits.

Ordinary product use costs nothing: viewing jobs and candidates, editing a brief by hand, building a process, adding a note, sending a message you wrote yourself. Only the AI work around those actions (drafting, rewriting, scoring) uses credits.

## No seats

Every plan includes as many members as you invite. Adding a teammate to a workspace doesn't add to your bill: only the AI work they run does.

## The \$100 to start

Signing up grants an organization \$100 in credits, once, applied when its first workspace is created. Extra workspaces don't get their own grant. See [Plans](/pricing/plans).

## See your balance

The sidebar shows a **Credits** chip: your organization's available credits, or, if the workspace you're in has a monthly cap, this month's spend against that cap. Anyone in the workspace can see the chip. Only organization owners and admins can click through it to the full **Usage & credits** page. See [Credits, top-ups and spend caps](/admin/credits).

<CardGroup cols={2}>
  <Card title="Plans" icon="layers" href="/pricing/plans">
    Starter, Growth, and Enterprise, and how the top-up bonus works.
  </Card>

  <Card title="What uses credits" icon="list-checks" href="/pricing/what-you-pay">
    Which actions are metered, the one fixed price, and how to see where credits went.
  </Card>
</CardGroup>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.