<p>It is possible to use <strong>DeepInfra</strong> too (since version 1.2132).</p>
<ul>
<li>Create an account on the <ahref="https://deepinfra.com/">DeepInfra website</a> and deploy a model, for example, <em>meta-llama/Llama-2-13b-chat-hf</em></li>
<li>In the integration settings enter the URI https://api.deepinfra.com/v1/openai, an API key and the model name</li>
<li>On the integration settings enter the URI https://api.deepinfra.com/v1/openai, an API key and the model name</li>
</ul>
<p>You can also use <strong>Mistral</strong>:</p>
<ul>
<li>Create an account <ahref="https://console.mistral.ai">via here</a></li>
<li>In the integration settings enter the URI https://api.mistral.ai/v1/, an API key and the model name</li>
<li>On the integration settings enter the URI https://api.mistral.ai/v1/, an API key and the model name</li>
</ul>
<p>You can also use <strong>OpenRouter</strong>:</p>
<ul>
<li>Create an account <ahref="https://openrouter.ai/">via here</a></li>
<li>On the integration settings enter the URI https://openrouter.ai/api/v1, an API key and a model name</li>
<li>Enter a maximum number of tokens; suggested value: 2000 (since version 1.2333)</li>
</ul>
<p><br></p>
<p><br></p>
<p>If you are looking for Google Gemini support, please see <ahref="#faq204">this FAQ</a>.</p>
<p>This feature is experimental and requires version 1.2053 or later for the GitHub version and version 1.2182 or later for the Play Store version.</p>