This week we announced Veyra-1, our new multimodal reasoning model. It plans, checks and corrects its own reasoning before it answers. Veyra One is the device we designed around it: a 62 g pebble that listens, sees and thinks with you, all day.
No apps, no screen to scroll. Press the orb, talk, and Veyra One answers out loud. It remembers what matters to you and forgets what you ask it to.

Real conversations with under 300 ms response time, in 38 languages.
Point it at a menu, a sign or a document to translate, explain or summarize it.
Everyday requests run on-device. The rest is processed in the EU and never used for training.
A week of typical use on a single charge. Charges wirelessly in 40 minutes.
Real prompts from our launch keynote, replayed on Veyra One.

| AI model | Veyra-1 Lite on-device + Veyra-1 in the EU cloud |
| Processor | Veyra NPU, 12 TOPS |
| Display | 1.4" circular AMOLED orb |
| Audio | 4-mic array, full-range speaker |
| Connectivity | Wi-Fi 6, Bluetooth 5.4, eSIM |
| Battery | 7 days typical, Qi2 wireless charging |
| Weight / colours | 62 g · Obsidian, Pearl |
Engineering prototype running Veyra-1 Lite on-device.
500 units built and tested by 120 beta users.
You are here. Pledges fund the first production run.
Manufacturing with our partner in Penang, Malaysia.
Numbered units ship first, worldwide.
Early Bird first, then the other tiers.
Hardware is hard. Our main risks are component supply (the NPU and the AMOLED orb) and certification timelines. We have secured components for the first 15,000 units and started CE/FCC certification in September. If anything slips, you'll hear it from us first, in the updates.