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

# What agents cost

> An agent is priced by the work it does on each row, not by the answer, so effort and web research move the number.

An agent is priced by the work it does on each row, not by the answer it gives. Effort, web research and added tools each move that price, and an estimate shows before you spend it.

<Frame caption="Rate chips beside Effort and Web research, before anything runs.">
  <img src="https://mintlify.s3.us-west-1.amazonaws.com/oneprofile/images/agents/what-agents-cost.png" alt="The agent editor's Effort picker and Web research switch, each with a credit chip showing what a row costs at that setting." />
</Frame>

## What moves the price

Three things move what one row costs:

* **Effort**, the level you pick or leave on **Auto**
* **Web research**, which adds its own surcharge when it is on
* An added server under **MCP servers**, billed at its own rate when web research is off

## See the estimate before you run

The price per row shows wherever you are about to spend it. That means beside each effort level, on the **Web research** switch, and on the button that commits a deploy.

<Steps>
  <Step title="Check the picker">
    Open **Effort** to see every level's price per row, or click **Compare effort levels** to see them side by side.
  </Step>

  <Step title="Check the deploy dialog">
    When you click **Add to table**, the button that commits your choice opens a menu naming each batch and its price before anything runs.
  </Step>
</Steps>

## Good to know

<AccordionGroup>
  <Accordion title="Why does the deploy dialog say the price might be higher?">
    An added MCP server is not visible to that estimate, so the number shown is a floor rather than a ceiling when one is attached.
  </Accordion>

  <Accordion title="Does turning off web research always lower the cost?">
    Usually, unless the agent has an MCP server added. Then the run still bills at that server's own rate rather than the model's plain rate.
  </Accordion>

  <Accordion title="Does training cost credits too?">
    Yes. Training runs the agent over your sample rows at the same per-row rate a deployed column would pay, and the estimate shows before you start.
  </Accordion>

  <Accordion title="Where do I see what an agent has actually spent?">
    Your overall spend, including agent runs, is on the [Credits and billing](/settings/credits-and-billing) page.
  </Accordion>
</AccordionGroup>

## Next steps

<CardGroup cols={2}>
  <Card title="Choose an effort level" icon="timer" href="/agents/choose-an-effort-level">
    Weigh cost against depth before you pick one.
  </Card>

  <Card title="Credits and billing" icon="credit-card" href="/settings/credits-and-billing">
    See your credit balance and the two ways to buy more.
  </Card>
</CardGroup>