Blog
Why we built Mylo
On ownership, agency, and building tools that actually belong to you
Every tool you use today treats you as a generic user. We built Mylo around a single belief — that your tools should know you, not just respond to you.
Why we chose Tauri over Electron
The case for native performance and a Rust backend
Smaller binaries, lower memory usage, and native performance. What we gained, what we gave up, and why it matters for a local-first AI app.
8 min readThe cognitive architecture behind Mylo's AI
Perceive, Reason, Decide, Plan, Validate
How a five-stage cognitive loop enables proactive behavior without losing user control. The theory and the implementation.
12 min readLocal-first AI: why your data belongs on your device
The cloud isn't wrong — it's just the wrong default
A case for privacy-first architecture in an AI-native world. Why we chose libSQL over cloud databases, and what that means for you.
7 min readRooms: how we build apps inside an app
React sandboxes, esbuild pipelines, and security boundaries
The technical story behind letting users build custom workspaces with natural language, rendered inside a secure iframe.
10 min readMulti-agent systems: what the research actually says
A survey of 14 papers on agent collaboration
Task delegation, multi-agent debate, ensemble routing. When multiple agents actually beat one, and when they don't.
15 min read