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

# Set up AI, credits and your own API key

> Choose how AI is paid for and which assistance your workspace can use.

<Note>
  This guide includes features awaiting release. If an option is not available in your workspace, contact StackShift Support.
</Note>

## What you can do

Choose how AI is paid for and which assistance your workspace can use. Where to find it: AI → AI settings; AI → Usage and costs

## Before you start

* An owner or administrator with permission to change AI settings.
* For your own key, an OpenAI account with an eligible model and provider billing.

## Steps

<Steps>
  <Step>
    Open AI settings and choose Managed or Use your own key under Provider mode.
  </Step>

  <Step>
    For Managed, review available workspace credits and included plan allowance. For your own key, select an offered OpenAI model and enter the key in the protected API key field.
  </Step>

  <Step>
    Choose the staff features you need: Staff copilot, Knowledge answers and Audio transcription. Configure customer replies separately.
  </Step>

  <Step>
    Review the displayed limits and policies, save, then run one small staff request. Inspect its result and Usage and costs.
  </Step>

  <Step>
    When credits are low, review the available top-up and plan options. If using your own key, check your provider account for its charges and limits.
  </Step>
</Steps>

## Example

A small team starts with managed credits for reply drafts, then chooses its own OpenAI key when it needs to manage provider billing directly.

## Things to know

* Included credits and prices come from your current plan; do not assume a fixed number of replies per credit. Requests vary in length and model cost.
* Using your own key bills the provider directly and does not bypass workspace permissions or feature limits.
* Stored keys are never displayed. Leave the field blank to keep the existing key; never paste keys into conversations.

## What happens next

<Check>
  AI uses the selected workspace provider settings, and usage records let you inspect what was consumed.
</Check>

## If something goes wrong

<Warning>
  * No model offered or provider unavailable: review Availability and the specific reason before changing settings.
  * Request rejected: check credit balance or provider billing, model access and workspace limits.
  * Setting missing: ask the workspace owner to check your role.
</Warning>

## Related guides

<CardGroup cols={2}>
  <Card title="Use AI to help your support team" href="/stackshift-support/staff-copilot">
    Draft replies, summarize conversations and research an answer before a person sends it.
  </Card>

  <Card title="Let AI answer customers" href="/stackshift-support/customer-ai">
    Enable customer-facing replies for a chosen scope, with rules for handing work to a person.
  </Card>

  <Card title="Choose a plan and understand your usage" href="/stackshift-support/billing-and-plans">
    Review prices, staff allowances, conversation usage, storage and AI credits before paying.
  </Card>

  <Card title="Support AI, media and handling reporting" href="/stackshift-support/ai-media-and-handling-reporting">
    Configure AI funding, assist your team, translate articles and understand usage.
  </Card>
</CardGroup>


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