The models behind the tools

Most AI aggregators hide what's running under the hood. We don't. Depending on your task, your request is routed to the model best suited to it, balancing quality, speed and cost.

For everyday chat we use fast, efficient models; for harder reasoning or final-quality creative work, we route to stronger ones. On each response we show you which model produced it, so you always know.

Free plans use automatic routing. Pro and Max let you choose the model yourself when you want control. We add and retire models over time; this page reflects what we currently use.

EU and UK traffic is always routed to compliant endpoints (Mistral or paid Google). We never route privacy-sensitive regions to non-compliant providers.