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

# Elephas AI, Own API Key or Local AI: Which Should You Use?

> Compare the three ways Elephas can answer: Elephas AI, your own API key, or Local AI on your computer. Where your data goes, cost, quality, speed and plans.

Elephas always searches your documents on your computer. What you choose is where the answer is written: by **Elephas AI**, included with your plan; by a provider you pay directly with **your own API key**; or by **Local AI**, a model running on your computer. You control where AI runs, and you can switch at any time in **Settings › AI**.

## Before you start

* Sign in, and have an Elephas trial or plan. All three options need one. See [Plans and pricing](/docs/plans-and-billing/plans-and-pricing).
* In a Team workspace, the Team Owner's AI policy decides what is allowed. See [Team AI policy](/docs/teams/team-ai-policy).

## Compare the options

| | Elephas AI | Your own API key | Local AI |
| - | - | - | - |
| **Label in Settings** | **Elephas AI (included with your plan)** | **My own AI account** | **Local (on this computer)** |
| **Where the answer is written** | A third-party model provider, through the Elephas gateway | The provider you choose, such as OpenAI, Anthropic or Google Gemini | A model on your computer |
| **What leaves your computer** | Your question, a few recent messages and the matching passages | The same, sent straight to your provider | Nothing goes to an AI provider |
| **Personal information protection** | On by default | On by default | Not needed |
| **Cost** | Included with your plan, up to your AI allowance | Your provider bills you for what you use | Included with your plan. No usage limit. |
| **Setup** | None, or a one-time activation on this device | Paste a key and pick a model | One download of about 1.2 to 5.3 GB |
| **Answers** | Elephas reads further into your documents when it needs to, and checks each answer against its sources | The same, with the model you pick | A smaller model answers in one pass. Best for focused questions. |
| **Speed** | Fast | Depends on your provider and model | Slower. Depends on your computer. |
| **Works without internet** | No | No | Yes, after a recent plan check |
| **Team workspaces** | Yes, under the Team's AI policy | No | Yes, if the Team's AI policy allows it |

Indexing happens on your computer. Always. Your files and your library are never uploaded to any AI, whichever option you choose. See [What data is sent to AI](/docs/privacy-and-security/what-data-is-sent-to-ai).

## Steps

<Steps>
  <Step title="Open Settings › AI">
    Press ⌘, on Mac or Ctrl+, on Windows, then click **AI**. Under **How Elephas answers**, each option is a card that says in one line where your question and sources go.
  </Step>

  <Step title="Click the option you want">
    * **Elephas AI (included with your plan)** works right away, or after a one-time **Activate Elephas AI** on this device. See [Elephas AI](/docs/ai-options/elephas-ai).
    * **My own AI account** opens your list of providers. Choosing an account switches to it. See [Use your own API key](/docs/ai-options/use-your-own-api-key).
    * **Local (on this computer)** asks to download a model the first time, for example **Download Qwen 4B, 2.6 GB?**. Until you click **Download**, your previous option stays in use. See [Run Local AI](/docs/ai-options/run-local-ai).
  </Step>

  <Step title="Check the answer">
    Ask a question. The chat bar names the AI that answered: **Answered by: Elephas AI**, **Answered by: OpenAI · GPT-5** for your own key, or **Answered by: This computer · Qwen 4B** for Local AI.
  </Step>
</Steps>

The steps are the same on Mac and Windows.

## How it works

### Which option to choose

* **Choose Elephas AI** if you want good answers with no setup and no separate bill.
* **Choose your own API key** if you already pay for an AI provider, want a specific model, or want your data to go straight to a provider you have an agreement with. Elephas is not in the path.
* **Choose Local AI** for documents that should not leave your computer at all, or for working without internet. Answers are slower and simpler.

A common setup is Local AI for client files and Elephas AI for everything else. The setting is a single choice, so switch it in **Settings › AI** when you change work.

### Switching never deletes anything

Switching away from your own key keeps the key in your system keychain. Switching away from Local AI keeps the model on your computer. Come back at any time without setting up again.

### The plan check for your own key and Local AI

Your own key and Local AI need an Elephas trial or plan. Elephas checks your plan online before using them. A passed check lasts up to 60 minutes and is cleared when you restart Elephas. The check sends only a device ID, never your key, questions or documents. If Elephas can't reach its service when a check is due, Ask says **Elephas could not confirm your plan right now. Check your connection and try again.** See [Use Elephas offline](/docs/privacy-and-security/use-elephas-offline).

### In a Team workspace

Team questions use Elephas AI under the Team's AI policy, or Local AI if the policy allows it. Your own API key is never used for Team content, even if it is your current setting. See [Team AI policy](/docs/teams/team-ai-policy).

## Troubleshooting

<AccordionGroup>
  <Accordion title="Local (on this computer) is greyed out">
    The card says **Not enough free memory on this computer.** Your computer can't run even the smallest local model. Use Elephas AI or your own key. See [Local AI troubleshooting](/docs/troubleshooting/local-ai-troubleshooting).
  </Accordion>

  <Accordion title="Ask says my own key and the on-device model are paused">
    Your trial or plan has ended, or you are signed out. Your key, models and Brains are unchanged. Choose a plan in **Settings › Plan & Billing**. See [When your plan ends](/docs/plans-and-billing/when-your-plan-ends).
  </Accordion>

  <Accordion title="My Elephas AI allowance is used up">
    Switch to your own API key or Local AI to keep asking, or wait for the reset date. See [AI allowance used up](/docs/troubleshooting/ai-allowance-used-up).
  </Accordion>
</AccordionGroup>

## Related articles

<CardGroup cols={2}>
  <Card title="What data is sent to AI" href="/docs/privacy-and-security/what-data-is-sent-to-ai" />

  <Card title="Elephas AI" href="/docs/ai-options/elephas-ai" />

  <Card title="Use your own API key" href="/docs/ai-options/use-your-own-api-key" />

  <Card title="Run Local AI" href="/docs/ai-options/run-local-ai" />
</CardGroup>


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