Work through maths, logic and coding questions in a private chat workspace. Ask follow-up questions and explore an answer step by step with your selected reasoning model. Check important answers independently. You can see the thinking. When the model replies, its chain-of-thought appears in a collapsible Thinking section above the final answer — expand it to follow and check the reasoning, or leave it collapsed to just read the conclusion. Both the thinking and the answer stream in as the model produces them, so you can watch it reason in real time and stop early once it's on the right track. Because step-by-step answers run longer, AI Reasoning automatically gives reasoning replies a larger length budget so they aren't cut off mid-thought. It runs on your device on a capable built-in model — your messages aren't uploaded with the local model, and it works offline once the model is downloaded. A free on-device reasoning model, DeepSeek R1 Distill 1.5B, ships ready to go and is selected and warmed up on first launch, with no account and no setup. Your conversations are saved encrypted on your device. AI Reasoning is deliberately focused: keep multiple saved conversations, find text within a chat, search across all your chats, tune the reply length and creativity per chat, and back up and restore your conversations. It's chat for reasoning, without the clutter — there are no tools, web browsing, document chat, folders or personas to get in the way. The Free tier is permanent, not a time-limited trial, and has no message or feature-use quota. It covers every feature but is limited to the included DeepSeek R1 Distill 1.5B on-device reasoning model. Pro adds larger local reasoning models and cloud providers; the tier gates only the model, so there are no Pro-only features.
Productivity · Windows · Free
