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

# Connect AI providers to power the Portal Agent

> Portals works out of the box with Pollinations at no cost. Connect OpenAI, Anthropic, Google Gemini, Groq, Ollama, or OpenRouter using your own API keys.

Portals works immediately without any setup — Pollinations is enabled by default and requires no API key. When you're ready to use a different model or provider, open **Settings → AI Configuration** and choose from the supported providers below. Your API keys are stored only on your device and are never sent to any Portals server.

<Note>
  API keys are saved locally on your device. Portals does not transmit your keys to any external server.
</Note>

<Tip>
  If you don't have any API keys yet, start with **Pollinations** (built-in, free) or **OpenRouter Free** to access a rotating selection of no-cost models.
</Tip>

## Supported providers

<AccordionGroup>
  <Accordion title="Pollinations — free, no key required">
    Pollinations is the default provider in Portals. It runs a free inference service and requires no sign-up or API key. Portal Agent uses it automatically the first time you open the app.

    **Default model:** selected automatically by Pollinations

    **How to use it:**
    No configuration needed. If you previously switched to another provider and want to return to Pollinations, open **Settings → AI Configuration**, select **Pollinations**, and save.
  </Accordion>

  <Accordion title="OpenAI">
    Use GPT-4o and other OpenAI models with your own API key.

    **Default model:** `gpt-4o`

    **Get an API key:**

    <Steps>
      <Step title="Go to platform.openai.com">
        Sign in or create an account at [platform.openai.com](https://platform.openai.com).
      </Step>

      <Step title="Create an API key">
        Navigate to **API keys** in the left sidebar and click **Create new secret key**. Copy the key — you won't be able to view it again.
      </Step>

      <Step title="Enter it in Portals">
        Open **Settings → AI Configuration**, select **OpenAI**, paste your key into the **API Key** field, and choose a model. Click **Save**.
      </Step>
    </Steps>

    You can optionally set an **Organization ID** and a custom **Base URL** if you are using a self-hosted or Azure OpenAI deployment.
  </Accordion>

  <Accordion title="Anthropic (Claude)">
    Use Claude models — including the claude-3-5-sonnet series — with your Anthropic API key.

    **Default model:** `claude-3-5-sonnet-20241022`

    **Get an API key:**

    <Steps>
      <Step title="Go to console.anthropic.com">
        Sign in or create an account at [console.anthropic.com](https://console.anthropic.com).
      </Step>

      <Step title="Create an API key">
        Navigate to **API Keys** and click **Create Key**. Copy the key.
      </Step>

      <Step title="Enter it in Portals">
        Open **Settings → AI Configuration**, select **Anthropic**, paste your key, choose a model, and click **Save**.
      </Step>
    </Steps>
  </Accordion>

  <Accordion title="Google Gemini">
    Use Gemini models via the Google AI Studio or Vertex AI API.

    **Default model:** `gemini-1.5-pro`

    **Get an API key:**

    <Steps>
      <Step title="Go to aistudio.google.com">
        Sign in at [aistudio.google.com](https://aistudio.google.com) with your Google account.
      </Step>

      <Step title="Create an API key">
        Click **Get API key** and follow the prompts. Copy the generated key.
      </Step>

      <Step title="Enter it in Portals">
        Open **Settings → AI Configuration**, select **Google Gemini**, paste your key, and click **Save**.
      </Step>
    </Steps>

    You can optionally provide a **Project ID** for Vertex AI deployments.
  </Accordion>

  <Accordion title="Groq">
    Groq provides fast inference for open-weight models at low latency.

    **Default model:** set during configuration

    **Get an API key:**

    <Steps>
      <Step title="Go to console.groq.com">
        Sign in or create an account at [console.groq.com](https://console.groq.com).
      </Step>

      <Step title="Create an API key">
        Navigate to **API Keys** and generate a new key. Copy it.
      </Step>

      <Step title="Enter it in Portals">
        Open **Settings → AI Configuration**, select **Groq**, paste your key, enter a model ID (for example `llama3-8b-8192`), and click **Save**.
      </Step>
    </Steps>
  </Accordion>

  <Accordion title="Ollama — local models">
    Ollama runs models entirely on your machine. No API key is required, and no data leaves your device.

    **Default model:** `llama3.2`
    **Default base URL:** `http://localhost:11434`

    **Set up Ollama:**

    <Steps>
      <Step title="Install Ollama">
        Download and install Ollama from [ollama.com](https://ollama.com). Start the Ollama service.
      </Step>

      <Step title="Pull a model">
        In your terminal, run `ollama pull llama3.2` (or any other supported model).
      </Step>

      <Step title="Enter it in Portals">
        Open **Settings → AI Configuration**, select **Ollama**, confirm the base URL is `http://localhost:11434`, enter the model name, and click **Save**.
      </Step>
    </Steps>

    <Note>
      Portals must be able to reach your Ollama server over HTTP. If you're running Portals in a browser on a remote machine, make sure the Ollama endpoint is accessible from that origin.
    </Note>
  </Accordion>

  <Accordion title="OpenRouter">
    OpenRouter routes requests to hundreds of models from multiple providers using a single API key.

    **Default model:** `anthropic/claude-3.5-sonnet`

    **Get an API key:**

    <Steps>
      <Step title="Go to openrouter.ai">
        Sign in or create an account at [openrouter.ai](https://openrouter.ai).
      </Step>

      <Step title="Create an API key">
        Navigate to **Keys** in your account settings and generate a new key.
      </Step>

      <Step title="Enter it in Portals">
        Open **Settings → AI Configuration**, select **OpenRouter**, paste your key, enter a model ID (for example `anthropic/claude-3.5-sonnet`), and click **Save**.
      </Step>
    </Steps>

    You can optionally provide a **Site URL** and **Site Name** that OpenRouter will include in request attribution.
  </Accordion>

  <Accordion title="OpenRouter Free">
    OpenRouter Free gives you access to a rotating selection of models at no cost. It requires an OpenRouter API key but uses only the free-tier models that OpenRouter makes available.

    **How it works:**

    Portals fetches the current list of available free models from OpenRouter and lets you select one. The available models change over time as OpenRouter updates its free tier.

    <Steps>
      <Step title="Get an OpenRouter API key">
        Follow the same steps as the OpenRouter provider above to create an account and generate a key at [openrouter.ai](https://openrouter.ai).
      </Step>

      <Step title="Select OpenRouter Free in Portals">
        Open **Settings → AI Configuration**, select **OpenRouter Free**, paste your key, and click **Fetch free models**.
      </Step>

      <Step title="Choose a model">
        Select from the list of available free models and click **Save**.
      </Step>
    </Steps>
  </Accordion>
</AccordionGroup>

## OpenAI-compatible providers

<Accordion title="Connect any OpenAI-compatible endpoint">
  If you run a self-hosted inference server — such as LM Studio, vLLM, or a custom deployment — that exposes an OpenAI-compatible API, you can connect it using the **OpenAI Compatible** provider type.

  **Required fields:**

  * **Base URL** — the root URL of your API, for example `https://api.example.com/v1`
  * **API Key** — your authentication token (required by most servers, even local ones)
  * **Model** — the model identifier to pass in requests

  You can also provide a **Display Name** to identify this provider in the UI.

  Open **Settings → AI Configuration**, select **OpenAI Compatible**, fill in the fields above, and click **Save**.
</Accordion>