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

# AI Providers (BYO key)

> Bring your own OpenAI, Anthropic, or Google Gemini key to power AI generation, page cloning, quizzes, and AI chat features inside ElasticFunnels.

## Overview

ElasticFunnels uses AI for several features — page generation, page cloning, copy rewriting, AI chat, quiz building, split-test analysis, and on-page interaction generation. You bring your own API keys for any of these providers, and EF routes each task to the right model.

This is a **bring-your-own-key** integration: you control billing, rate limits, model selection, and which provider you trust for each task.

## Supported providers

<CardGroup cols={3}>
  <Card title="OpenAI" icon="circle-nodes">
    GPT-4-class and GPT-5-class models, plus reasoning models (o-series). Used for general copywriting, page generation, AI chat, and split-test analysis.
  </Card>

  <Card title="Anthropic Claude" icon="brain">
    Claude 3.5 Sonnet, 3 Opus, 3 Haiku. Used for long-form copy, page cloning, and reasoning-heavy workflows.
  </Card>

  <Card title="Google Gemini" icon="atom">
    Gemini 2.5 Pro, 2.5 Flash, 2.0 Flash. Optimized for page-event generation, script-rule generation, and multimodal page cloning with reference images.
  </Card>
</CardGroup>

## What needs a key

| Feature                              | Provider used (default) | Notes                                                   |
| ------------------------------------ | ----------------------- | ------------------------------------------------------- |
| AI page generation                   | OpenAI or Anthropic     | Picks based on prompt length and your default           |
| Page cloning from URL                | Anthropic or Gemini     | Gemini is preferred when reference images are extracted |
| Page cloning with screenshots        | Gemini                  | Multimodal model required                               |
| AI Assistant inside the page builder | OpenAI                  | Chat + targeted edits                                   |
| Page event / interaction generation  | Gemini 2.5 Flash        | Cheap, fast, well-suited to script generation           |
| Quiz builder AI                      | OpenAI                  | Question + outcome generation                           |
| Copy rewriting (component panel)     | OpenAI                  | One-shot with the user's style notes                    |
| Split-test analysis                  | OpenAI or Anthropic     | Reads conversion data and writes a recommendation       |
| Email copy assist                    | OpenAI                  | Subject + body suggestions                              |

If a feature needs a provider you haven't connected, the UI shows a "Connect provider" prompt linking back to the integration form.

## Setup

<Steps>
  <Step title="Get an API key">
    * **OpenAI**: [platform.openai.com](https://platform.openai.com) → **API Keys** → **Create new secret key**.
    * **Anthropic**: [console.anthropic.com](https://console.anthropic.com) → **Settings** → **API Keys** → **Create Key**.
    * **Google Gemini**: [aistudio.google.com](https://aistudio.google.com) → **Get API key** → **Create API key in new project**.
  </Step>

  <Step title="Activate the integration in EF">
    Go to **Integrations** → **AI** and click the provider tab.
  </Step>

  <Step title="Paste the key">
    Paste the API key into the form and click **Save**. Keys are encrypted at rest.
  </Step>

  <Step title="(Optional) Override default models">
    Per-feature model overrides live on the integration form. Defaults are sensible — only override if you're cost-tuning or want a specific model family for a particular task.
  </Step>

  <Step title="Test">
    Open any AI surface (the page builder's AI assistant, or **Pages** → **Generate with AI**) and verify the generation runs without authentication errors.
  </Step>
</Steps>

<Tip>
  **Why BYO key?** AI calls can quickly become the largest line item in a SaaS bill. Bringing your own key keeps cost transparent (you see it on your provider invoice) and lets you negotiate enterprise pricing or use credits.
</Tip>

## Cost & rate limits

EF tracks token usage and approximate cost per request inside the AI logs. You can view per-feature spend under **Logs** → **AI Generations**. Rate limits and monthly quotas are enforced by the provider, not EF — adjust them in the provider's dashboard.

## Privacy

* API keys are encrypted at rest and never logged.
* Prompts are only sent to the provider associated with the feature you're using.
* ElasticFunnels does not train any models on your data.
* Whether the provider trains on your prompts is governed by your account settings with that provider — Anthropic and OpenAI both let you opt out at the account level.

## Troubleshooting

<AccordionGroup>
  <Accordion title="Authentication errors">
    Regenerate the key in the provider's dashboard, paste the new value into EF, and save. Old keys may have been revoked.
  </Accordion>

  <Accordion title="Rate-limit errors">
    The provider is throttling your account. Either request a quota increase, or switch to a cheaper / faster model in the integration form (e.g. Gemini 2.5 Flash for page-event generation).
  </Accordion>

  <Accordion title="Generation cuts off mid-output">
    The model hit its max output token limit. Try a longer-context model from the same provider, or break the task into smaller prompts (the AI assistant supports follow-ups).
  </Accordion>
</AccordionGroup>