Skip to main content
Back to Portfolio

Compact Notation for Grok-Powered Agent Systems

Aglyphs is a compact symbolic notation that lets you describe complex AI agent systems in a single line. Built for Grok and modern agent workflows.

Grok
Notation
Opensource
aglyphs logo

Aglyphs lets you describe complex AI agent architectures, teams, memory systems, and workflows in **one compact line** β€” perfect for docs, code comments, whiteboards, and GitHub READMEs.

Why We Need Better Notation

Most of us currently describe agent systems with long paragraphs, messy diagrams, or code. None of these scale well when you’re working with complex teams of agents.

Aglyphs solves this by giving us a clean, text-native way to express architectures that would normally take dozens of lines to explain.

Core Primitives

The system is built on a small set of primitives:

`𝔾` β€” Grok Agent

`⨁` β€” Agentic Operator (supervisor)

`𝕄` β€” Memory (`𝕄˒`, `𝕄ˑ`, `𝕄ᡛ`)

`⊚` β€” Reflection

`βŠ•` β€” Human-in-the-Loop

Examples:

Basic React Agent:

𝔾[Ξ›Λ£ βŠ— 𝕋] (βŠ™ β†’ Ξ› β†’ ⊚ β†’ β–·) ⟳

Grok as Central Operator:

⨁(𝔾) ⇑ {𝔸₁[research + 𝕄ᡛ], 𝔸₂[execution + 𝕋], 𝔸₃[analysis]}

Production Loop with Human Approval

𝔾[Ξ›Λ£] β†’ ⨁ β†’ (βŠ™ β†’ ⊚ β†’ βŠ• β†’ β–·) ⟳

Get Started

The full notation specification, comparison with other methods, and contribution guidelines are available in the repository:

β†’ github.com/rb-thompson/aglyphs

I’m excited to see what patterns the community creates with it. If you build something interesting using Aglyphs, feel free to open a PR or reach out.