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

# Connectors

> Connecting ChatGPT, Claude, Gemini, or Kimi with your own API key.

**Generatify > Settings > Connectors**

Connectors are how Generatify reaches an AI provider. Each one is a toggle and an API key. The whole tab is Pro, which is why every AI feature that calls a provider is also Pro.

<Info>
  A license unlocks the connectors. It does not include AI usage. You get a key from the provider, and the provider bills you for what you send them.
</Info>

## AI providers

<ParamField path="ChatGPT" type="toggle" default="off">
  Pro. OpenAI. Also the provider used for image generation.
</ParamField>

<ParamField path="OpenAI API Key" type="password">
  From [platform.openai.com](https://platform.openai.com/api-keys). Validated for format before saving.
</ParamField>

<ParamField path="Claude" type="toggle" default="off">
  Pro. Anthropic.
</ParamField>

<ParamField path="Anthropic API Key" type="password">
  From [console.anthropic.com](https://console.anthropic.com/).
</ParamField>

<ParamField path="Gemini" type="toggle" default="off">
  Pro. Google.
</ParamField>

<ParamField path="Gemini API Key" type="password">
  From [aistudio.google.com](https://aistudio.google.com/apikey).
</ParamField>

<ParamField path="Kimi" type="toggle" default="off">
  Pro. Moonshot AI.
</ParamField>

<ParamField path="Kimi API Key" type="password">
  From the Moonshot AI platform console.
</ParamField>

## Enabling more than one

You can connect all four. The model selector on **Generatify > Chatbot** then lists the most capable model from each, plus an **Auto** option that picks for you.

Connecting several is useful for comparing answers and for continuing when one provider has an outage. It does not reduce cost, since you pay whichever one you actually use.

<Card title="Which provider to choose" icon="scale-balanced" horizontal href="/generatify/guides/choosing-a-provider">
  What each is good at, and what a key costs to get.
</Card>

## Where your key is stored

Keys are saved in the `generatify_settings` option, in your database, alongside every other setting.

<Warning>
  Anyone with administrator access to the site, or with database access, can read the key. Treat it as a shared credential.

  Give each site its own key rather than reusing one, so a compromised site can be cut off without disturbing the others. Set a spending limit in the provider's dashboard, which is your only hard ceiling if a key does leak. Rotate keys when an administrator leaves.
</Warning>

## Verifying a connection

Go to **Generatify > Chatbot** and open the model selector at the bottom of the prompt box.

| What you see                       | What it means                                                 |
| ---------------------------------- | ------------------------------------------------------------- |
| Your provider listed               | The key was accepted                                          |
| **No AI Model Available**          | No connector is enabled, or no key is saved                   |
| The model listed but requests fail | The key is saved but rejected, out of credit, or rate limited |

For the third case, check the provider's dashboard rather than the plugin. Generatify passes their error back, and it will usually say which of the three it is.

## Requests that take a long time

The plugin allows up to three minutes for a provider to respond, which covers long generations on a slow model.

If your host enforces a shorter PHP execution limit, a long request can be cut off before the provider replies. Both the timeout and the execution limit are filterable:

<Columns cols={2}>
  <Card title="generatify_ai_request_timeout" icon="clock" href="/generatify/reference/hooks#generatify_ai_request_timeout">
    How long to wait for the provider.
  </Card>

  <Card title="generatify_ai_time_limit" icon="hourglass" href="/generatify/reference/hooks#generatify_ai_time_limit">
    The PHP execution limit for the request.
  </Card>
</Columns>