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

# Plans & Credits

> What each plan includes, what credits pay for, and what happens when you run out mid-build.

Everything on pre.dev — spec generation, build sprints, browser agent tasks — draws from one credit balance. Every plan refills that balance monthly.

<Note>
  [pre.dev/pricing](https://pre.dev/pricing) is the source of truth for current prices, credit amounts, and plan features. The numbers below reflect the pricing page at the time of writing.
</Note>

## Plans

| Plan           | Price    | Monthly credits | Adds                                                                                                                                        |
| -------------- | -------- | --------------- | ------------------------------------------------------------------------------------------------------------------------------------------- |
| **Free**       | \$0      | 100             | Everything you need to plan and build: architecture planning, autonomous coding and verification, deployment and preview, GitHub automation |
| **Plus**       | \$25/mo  | 200             | More credits                                                                                                                                |
| **Premium**    | \$49/mo  | 500             | [Pro Mode](/coding-agent/building/pro-mode)                                                                                                 |
| **Pro**        | \$199/mo | 2,500           | Deep Spec                                                                                                                                   |
| **Enterprise** | Custom   | Unlimited       | Architect API + MCP, dedicated solutions engineer                                                                                           |

A few feature notes, per the [pricing page](https://pre.dev/pricing):

* **[Pro Mode](/coding-agent/building/pro-mode)** — pin the strongest model to every phase of a build — is included from the **Premium** plan.
* **[Deep Spec](/coding-agent/specifications/fast-vs-deep)** — the granular, subtask-level specification — is included from the **Pro** plan. Fast Spec is available on every plan.
* Core building blocks — architecture planning, autonomous coding and verification agents, automatic deployment and preview, GitHub automation, and data-source connections — are on **every plan, including Free**.

## What credits pay for

<CardGroup cols={3}>
  <Card title="Spec generation" icon="compass" href="/coding-agent/specifications/fast-vs-deep">
    **Fast Spec:** \~5–10 credits, \~1 minute.
    **Deep Spec:** \~10–50 credits, \~3–5 minutes.
    Cost scales with project complexity.
  </Card>

  <Card title="Build work" icon="hammer" href="/coding-agent/building/build-modes">
    Sprints consume credits as the agent researches, codes, and verifies. Higher [effort levels](/coding-agent/building/build-modes#effort-level-how-deep-each-sprint-goes) do more model work and cost more.
  </Card>

  <Card title="Browser agent tasks" icon="browser" href="/browser-agents/overview">
    **1 credit = \$0.10**, billed per task with a **0.1-credit floor**. Failed tasks are free.
  </Card>
</CardGroup>

<Tip>
  Failed browser agent tasks aren't billed, but you need at least the 0.1-credit floor available for a task to start.
</Tip>

## Check your balance

Three ways, depending on where you are:

<CardGroup cols={3}>
  <Card title="In the workspace" icon="terminal" href="/coding-agent/workspace">
    Type `/balance` in the chat composer for your remaining credits.
  </Card>

  <Card title="On the web" icon="key" href="https://pre.dev/projects/key">
    [pre.dev/projects/key](https://pre.dev/projects/key) shows your balance alongside your API key.
  </Card>

  <Card title="From the API" icon="code" href="/architect-agent/api/credits-balance">
    `GET /credits-balance` returns your remaining credits programmatically.
  </Card>
</CardGroup>

## Per-project cost analytics

The workspace tracks what each project has spent, so you can see where credits actually go — which sprints, which phases, which features. Use it to decide where auto effort is fine and where a cheaper [low-effort pass](/coding-agent/building/build-modes) would do.

## What happens when credits run out mid-build

Running out of credits during a sprint doesn't lose your work:

<Steps>
  <Step title="The build pauses safely">
    The agent stops at a clean point. Nothing is discarded — the code, the plan, and the sprint's progress are all preserved.
  </Step>

  <Step title="You get an email">
    It includes the build's current progress, screenshots of the app so far, and a link to the live preview — so you can judge how far it got without opening the workspace.
  </Step>

  <Step title="Top up and resume">
    Add credits or upgrade at [pre.dev/pricing](https://pre.dev/pricing), then resume the build. It picks up where it paused.
  </Step>
</Steps>

## FAQ

<AccordionGroup>
  <Accordion title="Do unused credits roll over?">
    See [pre.dev/pricing](https://pre.dev/pricing) for the current policy on credit refills and rollover.
  </Accordion>

  <Accordion title="Which spec type should I budget for?">
    Fast Spec (\~5–10 credits) is the default and right for most projects. Deep Spec (\~10–50 credits) adds granular subtasks and is worth it for complex, production-bound builds. See [Fast vs Deep](/coding-agent/specifications/fast-vs-deep).
  </Accordion>

  <Accordion title="How far do credits go on browser agents?">
    At 1 credit = \$0.10 with a 0.1-credit floor per task, a credit covers up to \~10 simple tasks — and failed tasks cost nothing.
  </Accordion>

  <Accordion title="I need more than the Pro plan offers.">
    The Enterprise plan has unlimited credits, the Architect API + MCP, and a dedicated solutions engineer — [contact us via pricing](https://pre.dev/pricing).
  </Accordion>
</AccordionGroup>