SyxonAI‑native OS that preserves attention
Your knowledge, verifiable and alive — for students, founders, and researchers who think in notes. Local‑first and private‑by‑default.
Problem → Why now
- Chat made AI accessible but fractured work; threads/tabs don’t compound knowledge.
- On‑device models crossed usability thresholds; privacy/regulatory tailwinds.
- No mainstream substrate that is attribution‑first, democratic, and verifiable.
Solution → What
- Views on a verifiable knowledge substrate (ions/connections).
- Local‑first by default; cloud assist when you choose; explainable outcomes.
- Fractal navigation that preserves attention and compounds memory.
Time‑to‑first‑ion ≤ 5 min
Offline capture→persist→search
Attribution by design
Moat → Differentiators
- Attention‑preserving UX (map‑first, predictable flows; safe UI envelopes).
- Local‑first intelligence with clear local/cloud posture.
- Democratic governance: proposals, merges, typed/weighted relations.
- Verifiable knowledge: sources, uncertainty bands, cluster summaries.
- Substrate + tools: CLI, renderer, policy guards; multimodal path (2D+audio → 3D).
GTM → Wedges
- Students: study and research trails, exam prep, and campus programs.
- Founders: decision briefs, project hubs, and builder cohorts.
- Researchers: literature/research trails plus Notion/Obsidian import, browser capture, and CLI.
Business model
- Free local (private, offline); Pro (sync, collab, advanced models); Enterprise (SSO, compliance).
- Marketplace for experience packs/tools; revenue share.
Roadmap
- M1: 2D‑first + audio capture (Web Speech) + basic TTS. Policy‑guarded renderer.
- M2: 3D lens (scene3d) + optional local Whisper; streaming later via WebRTC/WebTransport.
Team & IP
- Founder: Zauh Jariwala. Prior: Braille Sense (PFH 2023), Ethical Attractors (ISDC 2025), Noise Sink (eResearch 2024).
- Provisional patent (IP Australia, 4 Aug 2025): Adaptive privacy‑preserving interface with fractal navigation.