Architecture
A local-first knowledge system, not one more note pile.
ATLAS:Face organizes your work as an intuitive, filesystem-backed hierarchy: Vaults contain Projects, projects contain Topics, and topics contain markdown documents, cards, media resources, links, review metadata, layout metadata, and source progress.
Core Objects
- Vaults are top-level folders for an entire body of knowledge or client domain.
- Projects group a course, research path, topic domain, or creative endeavor.
- Topics are the main working unit. A topic can be a markdown document, a card set, a source session, or all of them at once.
- Cards are interactive knowledge blocks with titles, content, SRS state, links, and code blocks.
- Resources attach PDFs, EPUBs, and videos directly to the topics they inform.
How It Feels in Practice
Start with a raw PDF source, a curious question, or a blank topic. Read or watch inside ATLAS:Face, extract what matters into cards and notes, connect related ideas on a spatial canvas, review them via spaced repetition, and publish clean versions when they are ready to share.
Sources
PDFs, EPUBs, and YouTube videos become active working material.
The Resources panel is the starting point for source-first study. It manages PDFs, EPUB books, and video resources so your reading progress, annotations, extracted cards, and comprehension checks stay anchored to the topic instead of fragmented across external apps.
PDF Reader
Open and annotate PDF documents, track page progress, draw with pens, search text, and keep notes connected to the source.
EPUB Reader
Read EPUB books with chapter navigation and inline card creation. The reader uses focused excerpts around your current location for targeted quiz generation.
YouTube Player
Study videos with transcript-aware AI tools. Generate timestamped questions, keep notes beside playback, and turn lectures into reviewable card decks.
Cards & Recall
Cards are movable, linkable units of understanding.
ATLAS:Face cards are much more than simple flashcards. They can be arranged on a spatial canvas, edited in place, linked with typed relationships, turned into spaced repetition review decks, presented in full-screen slideshows, and exported as standard markdown.
related, prerequisite, extends, contradicts, and supports.Markdown System
Topics are plain markdown documents with card-aware structure.
ATLAS:Face stores topic documents as plain markdown on disk. This gives you total ownership, git versioning, and scriptability while powering visual cards, layouts, SRS metadata, and graph links. Content under each level-two heading (##) automatically maps to a card.
---
title: Spanish Vocabulary
---
Opening notes before the first card can become topic preamble.
## Hola
Hello
## Gracias
Thank you
## Code example
```js
console.log("cards can preserve fenced code blocks");
```
ATLAS:Face preserves card metadata separately from the markdown document. Editing the document in external editors (like Obsidian or VS Code) rebuilds cards while preserving SRS review state, grid positions, and relations.
Journals & Widgets
Your journal can pull live context from the system.
The Journal is a persistent workspace panel for daily notes, session reflections, research planning, and reusable templates. It supports live widget tokens that automatically expand into snapshots of routines, tasks, focus areas, or daily progress.
Widget Tokens
Use these tokens in a journal template or daily note to embed real-time system state:
{{TODAY}}
{{ROUTINE}}
{{TASKS}}
{{FOCUS}}
What Widgets Include
- Today Snapshot: routine, tasks, and focus combined in one cohesive block.
- Routine Snapshot: scheduled time blocks and live completion statuses.
- Task Snapshot: active backlog tasks and tasks completed today.
- Focus Snapshot: active focus themes, prompts, and study goals.
Routines, Tasks, & Focus
Planning lives directly beside knowledge work.
ATLAS:Face integrates an intentional planning layer: knowledge compounds only when you regularly revisit it. The routine system, task panel, daily focus goals, and activity timeline give you structured momentum without leaving your workspace.
Routine Planner
Create weekly templates and daily flows with blocks for rituals, deep study, breaks, creative work, and review sessions. Blocks track timing, recurrence, priority, and linked topics.
Task Manager
Track tasks, recurring habits, active checklists, and completion history. Checked task states can automatically mirror into daily notes.
Daily Focus & Goals
Set daily focus themes, study reminders, review streaks, and mastery targets to keep learning oriented toward genuine depth.
Spatial Navigation
See ideas spatially when linear lists stop helping.
The Knowledge Universe and Canvas View provide spatial clarity for complex domains. Freeform canvas lets you arrange cards visually, while the Knowledge Universe renders your concepts as a 3D interactive force-directed graph.
AI Workbench
AI as precision workbench tools, not a generic chatbot.
AI features in ATLAS:Face are designed as surgical tools for learning: generate cards from sources, ask targeted questions about a specific card, quiz yourself on video lectures, deepen understanding, split dense cards, expand thin notes, and generate mock practice exams.
Generation
Convert raw text, PDFs, and video transcripts into structured flashcards and test questions with one click.
Contextual Dialogue
Card AI Chat lets you interrogate a specific card or concept in isolation without cluttering your global workspace.
Refinement & Exams
Think Deeper, Card Splitter, Card Expander, and the Exam Generator transform rough notes into rigorous active recall drills.
Publishing
Publish polished topics to the web with zero lock-in.
Turn any topic into an interactive, standalone web page. The local-first export creates self-contained HTML/JS bundles, while GitHub Pages publishing pushes structured study notes directly to your personal domain with interactive review features preserved.
Marketplaces
Install community knowledge packs and extend the app.
ATLAS:Face features two connected ecosystems: the Community Marketplace for shared topic packs, study guides, and subject templates, and the Plugin Marketplace for installable toolkits, visualizations, and workflow automations.
Community Content
Browse, install, and share curated study decks, exam preps, and research vaults across science, engineering, languages, medicine, and the humanities.
Plugin Ecosystem
Install plugins that add specialized panels, custom visualizations, external sync pipelines, timers, and specialized hotkeys.
CLI Reference
The af CLI is engineered for power users and AI agents.
The CLI can manipulate vaults, projects, topics, cards, links, resources, journal entries, backups, imports, exports, and SRS sessions. It provides full automation for terminal power users, cron scripts, and autonomous coding agents like Antigravity, Codex, and Claude Code.
cd cli
npm install
npm link
# Inspect status
af status
# Hierarchy management
af vault create "Research"
af project create "Machine Learning" --vault "Research"
af topic create "Transformers" --project "Machine Learning"
# Card operations
af card create --topic "Transformers" --front "What is self-attention?" --back "A mechanism relating positions of a sequence."
# Spaced repetition study
af study start --project "Machine Learning"
af study rate <card-id> --topic "Transformers" --rating 2
Primary Command Groups
af status— checks app status and IPC connection.af data— exports, imports, and manages backups.af vault / project / topic— organizes knowledge hierarchy.af card— creates, updates, links, and lists cards.af study— runs headless SRS sessions and rating.af journal— appends daily notes and snapshots.
Agent-Friendly Architecture
- Every command supports
--jsonfor clean parsing. - Direct file-access fallback when desktop UI is closed.
AF_USER_DATAenvironment variable for sandboxing.- Batch import via stdin or file:
af topic import-md.
Plugin SDK
Build custom plugins and tools for ATLAS:Face.
The Plugin SDK gives you a structured API to build custom panels, listen to app lifecycle events, register global hotkeys, access local storage, and create tailored review drills.
my-plugin/
manifest.json
plugin.js
README.md
styles.css
Customization
Configure profiles, shortcuts, and sync workflows.
The feature profile selector lets you scale ATLAS:Face from a distraction-free minimalist reader to an all-in-one knowledge workstation.
Get Started
Download ATLAS:Face for Your Platform
The application is 100% free, local-first, and runs natively across Windows, macOS, and Linux.
Start small, then let the system compound.
Open one source, create a few cards, write raw notes, and let ATLAS:Face connect the rest over time. The power tools are there whenever you need them.
Download ATLAS:Face