System Definition
What this system is, what it isn't, and why it's not just another chat UI or a single-site userscript.
Documentation
System Reference
Not a feature list — this is the full reference for understanding how the system works, why it exists, and where its boundaries are.
Guide
Understand the runtime, systematically
If the homepage answers “what can it do,” the docs answer “why does it exist, how does it work, and where are the edges.”
Design Goals
Why upgrade an AI chat page into an execution system, rather than staying at the content-generation and single-turn conversation level.
Overall Architecture
From the entry layer, protocol layer, runtime, browser capability layer to tool ecosystem — the full execution chain.
Protocol Specification
ΩHERE, batch processing, high-fidelity transport, and complex payload handling — why they improve execution reliability.
Runtime Mechanics
How synchronous execution, async tasks, background runs, state tracking, recovery, and result feedback work together.
Browser Capabilities
Page reading, same-origin execution, tab routing, screenshots, and web context utilization within the system.
Tools & Ecosystem
How local tools, server-side tools, MCP hub, and external capability pools all plug into the same runtime logic.
Boundaries & Risks
Site differences, stability boundaries, host dependencies, scope limitations, and what's not ready for production.