Introduction to ATMOS
Relevant source files
The following files were used as context for generating this wiki page:
ATMOS is a Deepmind-style AI-first workspace ecosystem designed to provide a high-performance environment for developers. Built with a modern tech stack featuring Next.js 16, React 19, Rust, and Tauri, the system integrates a layered monorepo architecture to manage infrastructure, core engines, and business logic efficiently.
The project emphasizes a documentation-heavy and agent-friendly design, ensuring that both human developers and AI agents can navigate the codebase with high clarity. It features a robust Documentation System and a specialized Project Wiki module that transforms source code analysis into structured, research-grade documentation.
Sources: README.md:1-5, DOCUMENTATION.md:1-10, skills/project-wiki/README.md:1-10
System Architecture
The ATMOS ecosystem follows a strictly layered L1 -> L2 -> L3 -> App design pattern. This vertical isolation ensures that infrastructure concerns are separated from business logic and application entry points.
Backend Layers (Rust)
The backend is implemented in Rust to provide performance and safety, organized into four primary crates:
- L1: Infrastructure: Handles low-level concerns including Database management (SeaORM), the WebSocket Engine, caching, and background jobs.
- L2: Core Engine: Provides technical capabilities such as PTY (Pseudo-Terminal), Git integration, File System (FS) watching, and Tmux session management.
- L3: Core Service: Implements the high-level business logic and domain-specific rules.
- Agent Integration: A specialized vertical module for ACP (Agent Control Plane) clients, agent management, and external service integrations.
Frontend & Shared Packages
The frontend environment is split between web and desktop targets, supported by shared utility packages:
- Web App: A Next.js 16 application.
- Desktop App: A cross-platform application built with Tauri 2.0.
- Shared UI: A component library based on shadcn/ui.
- Shared Utilities: Common hooks and utility functions.
Sources: README.md:7-25, apps/landing/README.md:10-15
Architectural Flow Diagram
The following diagram illustrates the relationship between the different layers and applications within the ATMOS monorepo.
Sources: README.md:7-25
Documentation System Architecture
ATMOS utilizes a "Progressive Disclosure" documentation strategy. The system is designed so that users (and agents) only encounter information as they need it, moving from 60-second overviews to deep architectural dives.
Documentation Layers
| Layer | Type | Purpose | Target Audience |
|---|---|---|---|
| L1 | Entry Point | High-level project and agent navigation | Everyone / AI Agents |
| L2 | Deep Dive | Architecture, Tech Specs, and Decisions (ADR) | Architects / PMs |
| L3 | Component | Task-oriented work instructions | Developers |
| L4 | Code Structure | Detailed file/folder explanations | Deep Divers |
Sources: DOCUMENTATION.md:12-40
Hierarchy Visualization
This diagram shows how documentation is distributed across the physical file system of the project.
Sources: DOCUMENTATION.md:42-80
Project Wiki System
A core feature of the ATMOS workspace is the Project Wiki, a research-grade documentation generator. Unlike traditional generators, it performs deep source code analysis to explain the "why" and "how" of implementation rather than just summarizing README files.
Wiki Integration and Workflow
The Wiki is integrated into the "CenterStage" UI component as a fixed tab. It uses a 3-column viewer consisting of a navigation tree, markdown content renderer, and a table of contents (TOC).
Content Standards
Every article generated within the Wiki must adhere to strict quality metrics to ensure technical depth.
| Metric | Getting Started | Deep Dive / Specify |
|---|---|---|
| Min. Word Count | 800+ | 1500+ |
| Source References | 3+ Files | 5+ Files |
| Mermaid Diagrams | 2+ | 3+ |
| H2 Sections | 4+ | 6+ |
Sources: skills/project-wiki/README.md:66-75, specs/tech/project-wiki.md:130-140
Summary
The ATMOS platform is a highly structured, AI-ready workspace ecosystem. By leveraging a layered Rust backend and a modern React-based frontend, it provides the performance necessary for complex development tasks. Its standout feature is a deeply integrated documentation and wiki system that ensures technical knowledge is derived directly from the source code, maintaining high accuracy for both human contributors and automated agents.
Sources: README.md:1-5, DOCUMENTATION.md:150-160