Two models, one tap
Switch between Raven Flash and Raven Max inside a conversation. Fast by default; deep when it counts.
Raven app
Chat with the Raven model family from any device. Sign in with Google or Apple, pick your model, attach an image, and pick up the thread anywhere.
Switch between Raven Flash and Raven Max inside a conversation. Fast by default; deep when it counts.
Responses stream as they form. No spinners staring contests; the words arrive while the model works.
Raven streams a reasoning channel rendered as a collapsible thinking block. Watch the plan; keep it closed when you don't need it.
Attach photos, screenshots, and diagrams. Raven reads images alongside your text and answers about what it actually sees.
No new password to forget. Your plan, usage, and API keys live on one account.
Web today. Native apps for macOS, Windows, iPhone, and Android are landing through October.
Try it now
No download, no wait: sign in and start talking to Raven in your browser. Your conversations are yours; we don't log their content.
Open chat.dipoleml.com// a Raven conversation, abridged
You: Summarize this whiteboard photo and list the three risks we flagged.
Raven: The board covers three workstreams. The flagged risks are: 1) the launch date depends on app review, 2) billing edge cases around annual upgrades, 3) the data migration window overlaps with the holiday freeze.
Downloads
The web chat is live today. Desktop and mobile versions of the app are rolling out now.
Windows 10 and later
Apple silicon and Intel
iOS 17 and later
Coming October 15, 2026
Android 12 and later
Coming October 15, 2026Meanwhile, everything works in the browser at chat.dipoleml.com, and subscriptions carry across every device.