One anchor. Four adjacent.
We anchor in consumer and commerce, where multimodal capture and on-device inference unlock products that feel personal and don't bleed cloud spend. The same primitives carry into healthcare, legal, field service, and education.
Consumer & Commerce
Multimodal assistants embedded in iOS and Android products. Camera-to-action capture, voice-first cook and shopping flows, on-device language models that don't bill per request. Personal by default — your users' content, photos, and prompts never leave the device.
- → RecipeGuide — multimodal capture on-device
- → Apple Vision OCR + on-device LLM stack
- → Voice-first cook and shopping flows
Healthcare & Clinical
Privacy-first agents that handle PHI on-device — transcription, action-item extraction, clinical workflows. HIPAA-compatible by architecture, not by promise.
Legal
Document review, deposition summarization, contract diffing — all on-device for attorney-client privilege.
Field Service
Voice-first agents for technicians. Photo-to-pricing, parts lookup, hands-free workflows on-device.
Education
Reading-standards-aware adaptive content. Comprehension checks. Curriculum-aligned generation.