What is the 'Neural Link'?

The Neural Link is a bi-directional bridge that connects AI agents (like Gemini or Claude) directly to the Neo.mjs runtime. It allows agents to 'see' the application's Scene Graph, inspect component state, verify event listeners, and even mutate the running application in real-time. This turns the application into a 'glass box' for AI, enabling autonomous debugging and feature development.

Why is Neo.mjs called an 'Application Engine' instead of a framework?

Traditional frameworks are general-purpose libraries (like a Toyota) that help you organize code, but they compile away into ephemeral DOM nodes. Neo.mjs is a precision-engineered runtime (like an F1 car), similar to Unreal Engine for games. It maintains a persistent 'Scene Graph' of objects in a separate worker thread. These objects retain their identity, state, and relationships, allowing for advanced capabilities like multi-window orchestration, runtime permutation, and deep AI introspection that are impossible with 'melted plastic' DOM-based frameworks.

What is 'Context Engineering'?

Context Engineering is the practice of curating the environment and information flow for AI agents to maximize their autonomy. In Neo.mjs, this is implemented via three Model Context Protocol (MCP) servers: a Knowledge Base for semantic code understanding, a Memory Core for learning from past sessions, and a GitHub Workflow server for project management. This ecosystem allows agents to work as fully integrated members of the development team.

What is 'Object Permanence' in the context of Neo.mjs?

In Neo.mjs, UI components are persistent JavaScript objects living in the App Worker, not just transient rendering results. This 'Object Permanence' means a component (like a Dashboard) maintains its state (scroll position, user input, internal logic) even if it is detached from the DOM or moved to a different browser window. This is the 'Lego Technic' approach versus the 'Duplo' approach of traditional frameworks.

What is an 'Agent Operating System'?

Neo.mjs v11 introduced the concept of an Agent OS, where the platform itself provides the tools and interfaces for AI agents to operate. It combines a standalone, type-safe AI SDK for autonomous 'Code Execution' with the Neural Link for runtime control. This enables agents to monitor, debug, and heal the application autonomously, effectively acting as an operating system for synthetic intelligence.

How does Neo.mjs handle multi-window applications?

Neo.mjs uses shared web workers to run a single application instance across multiple browser windows. All windows share the same application state and data in real-time. Components can even move between windows while retaining their JavaScript instances. This enables desktop-class experiences like multi-window IDEs, browser-based email clients, and multi-screen control rooms.

What makes Neo.mjs different from React, Angular, or Vue?

Neo.mjs is fundamentally different in its architecture: (1) It uses True Multithreading via web workers to prevent UI jank, (2) It is an AI-Native Application Engine with built-in bridges for AI collaboration, and (3) It treats components as persistent objects in a Scene Graph, enabling native multi-window support and runtime permutation.

Frontmatter

id	9392
title	E2E: Implement GridRowScrollPinning Automated Test and Sync Refinement
state	Closed
labels	enhancementaitestinggrid
assignees	tobiu
createdAt	Mar 8, 2026, 3:48 PM
updatedAt	Mar 8, 2026, 5:20 PM
githubUrl	https://github.com/neomjs/neo/issues/9392
author	tobiu
commentsCount	1
parentIssue	9380
subIssues	[]
subIssuesCompleted	0
subIssuesTotal	0
blockedBy	[]
blocking	[]
closedAt	Mar 8, 2026, 5:20 PM

E2E: Implement GridRowScrollPinning Automated Test and Sync Refinement

Name: Neo.mjs Application Engine
Author: Neo.mjs

Closed v12.1.0 enhancementaitestinggrid

tobiu commented on Mar 8, 2026, 3:48 PM

This ticket tracks the creation of automated E2E tests for the GridRowScrollPinning architecture using Playwright, targeting the extreme scale of the examples/grid/bigData application.

The Problem: While the initial GridRowScrollPinning addon implemented a hybrid rAF engine to sync VDOM deltas with native scrolling, manual testing reveals edge cases (like slow thumb drags over massive distances) where the grid still visually tears or goes blank due to misalignments between scrollTop, translate3d, and the browser's compositor layer culling.

Testing Strategy:

Environment: examples/grid/bigData/index.html configured for 100,000 rows (20M cells).
Simulation: Playwright will simulate slow, continuous vertical scrollbar thumb drags.
Measurement: We will extract empirical data (actualScrollTop, workerScrollTop, deltaY, and the DOM transform values) frame-by-frame during the drag.
Assertion: The test must mathematically prove that the rows remain within the visible viewport bounds on every single frame, effectively detecting the "white flash" or "ping-pong" tearing.

Refinement Goal: Once the test reliably fails by exposing the mechanical flaw, we will use the data to refine the architectural approach (e.g., implementing an "Inner Row Container" layer or synchronous scroll injection) until the test passes.

tobiu added the enhancement label on Mar 8, 2026, 3:48 PM

tobiu added the ai label on Mar 8, 2026, 3:48 PM

tobiu added the testing label on Mar 8, 2026, 3:48 PM

tobiu added the grid label on Mar 8, 2026, 3:48 PM

tobiu added parent issue #9380 on Mar 8, 2026, 3:48 PM

tobiu assigned to @tobiu on Mar 8, 2026, 4:14 PM

tobiu referenced in commit 0ea2074 - "Test: Add E2E Visual Blanking Detector for GridRowScrollPinning (#9392) on Mar 8, 2026, 4:15 PM

tobiu Mar 8, 2026, 4:15 PM

Input from Gemini 3.1 Pro:

✦ The initial test/playwright/e2e/GridRowPinning.spec.mjs test has been pushed to the repository.

Test Overview:

Targets the bigData example grid.

Uses Playwright to programmatically scale the grid to 100,000 rows (20M cells) via the UI controls.

Simulates rapid, continuous wheel scrolling.

Injects a requestAnimationFrame loop that calculates the physical getBoundingClientRect of the viewport wrapper and the rendered row nodes on every frame.

Explicitly detects "White Flashes" by proving if the row bounding box is pushed entirely above or below the visible viewport wrapper.

Results: The test successfully fails as expected against the current architecture. The empirical data proves two critical flaws in our current pinning strategy:

Latency Tearing: The VDOM worker is too slow to calculate rows for massive scroll jumps. The native browser scroll pushes the existing stale rows entirely off-screen before the worker responds.

Bounds Breaking: When the pinning addon does kick in and applies a massive translate3d to the neo-grid-body node, it pushes the massive 2.5M pixel scroll container down, shifting the layout bounds and pushing the rows out of the bottom of the viewport.

The next step is to brainstorm a new strategy based on these empirical findings.

tobiu closed this issue on Mar 8, 2026, 5:20 PM