The voice layer for frontier AI.
Tap once and speak your mind. Riffie never cuts you off. Pause to think, keep going, finish when you're ready. Tap again to send, and you're hands-free as Riffie automatically plays the response. Full controls when you want them: scrub, change speed, share.
Pick the one that matches what you're trying to get done.
A definitive call without the clutter of overthinking. No options matrix, no qualifiers. Just a confident recommendation you can put into action.
Pressure-test your ideas like you would with a sharp friend. Push past easy answers, weak assumptions, and vague thinking. Strengthen your thinking.
A jazz musician's "yes, and" instinct, applied to your thinking. The half-baked thought you almost dropped might be the unlock two steps later. Cross-domain connections nobody asked for. You leave with more possibilities than you walked in with.
Warm, unhurried, with a light Stoic sensibility. Reflection without the therapy framing. The benefits compound: clearer thinking, less rumination, better sleep.
Opinions, tangents, dissecting trivial things for twenty minutes if that's the mood. Two people riffing, no agenda. Sometimes the rambling itself gives you what you need.
Hand over a topic that's hard to grasp — gaslighting, quantum computing, or why do cats love knocking things over. An explanation that builds understanding, piece by piece, like the best teachers do.
No persona layered on top — just the model's defaults coming through.
A model that fits the work, a voice that fits the moment.
Same chat, same memory — different model. Try one for a week, switch when the work changes.
Reads between the lines, like talking to a well-read friend.
Clear, direct, and reliable.
Fast and to the point. The most efficient of the three.
Each persona ships with a default voice tuned to its role. Override anytime.
More to choose from in the app.
Talk things through wherever you are, in an app built for spoken thinking.
Other voice modes water down the model and tuck the real response behind a mode switch. Riffie gives you the same depth you'd get typing, read aloud automatically, with the transcript right there.
A lot of our best thinking shows up on walks, drives, between calls. Riffie meets you there. Record a thought as long as you need, hear the response back, hands free for the wheel, the leash, the coffee.
A short note when the public version is ready. No newsletter, no marketing list.
We'll only email you once, at launch. Reply to opt out anytime.
More in the app, in Settings, or by email.
A voice-first AI app for iPhone and iPad. Tap once and speak as long as you want. Riffie replies in audio, with the full transcript right there. Built for thinking out loud, hands free.
There's a free tier with a small weekly budget — enough to get a feel for it. Pro is a paid subscription that unlocks a generous weekly budget for daily use. New accounts get a free trial at the Pro budget so you can try it before deciding.
iPhone and iPad, running iOS 26 or newer. No Android, no web app — at least not yet.
Not to try it. After the free trial ends, signing in is required so your conversations sync across devices and your subscription follows you. Sign in with Apple, Google, or email.
Riffie does not sell your data, run ads, or track you across apps. Speech transcription happens on-device. Conversations sync to your private iCloud, not to a Riffie server. Full details on the privacy page.
Yes. Pick the model that fits the work in Settings, and the voice that fits the moment per persona. Same conversation, same memory, different model — switch anytime.
Because that's what it's for — a place to riff out loud, with an AI that doesn't cut you off.