§ 01 — INTRODUCTION

Turn your agent into your mentor.

Odyss is the MCP layer that gives LLMs the structure, context and memory they need to guide real learning — mapped to your goals, skills and progress.

[ GOALS ] + [ KNOWLEDGE MAP ] + [ MEMORY ] = [ BETTER GUIDANCE ]

NOT GENERIC ANSWERS. NOT RANDOM ROADMAPS. — JUST THE NEXT RIGHT STEP.

Join the waitlist → See how it works

§ 02 — FROM GOALS TO MASTERY

A simple path. A structured system.

The same five moves, every conversation — so the LLM always knows where you are and where you're going.

1

Define goal

FIG. 03

Clarify the outcome and why it matters.

→ GOAL
SET

2

Map capabilities

FIG. 04

Break the goal into capabilities and skills.

→ CAP
MAP

3

Find the frontier

FIG. 05

Identify what you know and what's next.

→ FRONTIER
SET

4

Recommend next step

FIG. 06

Surface the highest-impact next step.

→ NEXT
STEP

5

Validate mastery

FIG. 07

Assess, reflect, and lock in the learning.

↳ ALREADY MAPPED EVERY NODE LIVE — every edge weighted.

1,249

Careers & roles

100% of O*NET plus 62 emerging — from welder to ML engineer.

4,204

Capabilities

Canonical, deduplicated clusters — things you can prove you can do.

25,974

Knowledge atoms

Atomic units, roughly an hour each, every one a learnable step.

179K

Mapped edges

Requires, assists, contributes-to, required-by — the wiring that makes the map walkable.

§ 03 — HOW THE KNOWLEDGE MAP WORKS

How the knowledge map works.

A STRUCTURED GRAPH THAT TURNS GENERIC
LLM ANSWERS INTO PERSONALIZED GUIDANCE.

9:41● ●●●

←

AI Tutor

Powered by Odyss

•••

I want to train my own models but the math feels overwhelming — what should I work on next?

9:41 AM

Next-best topic · 94% match

Backpropagation

You've already nailed Python ✓ and Linear Algebra ✓ — backprop is the missing piece. It's how every neural net learns; once it clicks, autograd, optimizers, and modern architectures all fall into place.

✓ Python ✓ Linear Algebra → Backprop

9:41 AM

Why this and not MLOps?

9:42 AM

Ranked by goal impact

Backprop94

MLOps58

Distrib. Train42

Highest return on your next 2 hours.

9:42 AM

Ask anything… ↑

↳ FIG. 02 — PERSONALIZED GUIDANCE IN CHAT

USER ASK →

ODYSS MCP

graph traversal

← NEXT STEP

↳ FIG. 03 — LIVE TRAVERSAL · GOAL → CAPABILITY → KNOWLEDGE

Learner state

Mastered

In progress

Not started

Live= Odyss walks the graph for every question — pulling in goal, mastered skills, and what's just past your edge.

You= mastered, in-progress and untouched topics, remembered across every conversation.

Prerequisite-aware graph.

Understands what unlocks what.

Persistent learner model.

Remembers your skills & progress.

Goal-based ranking.

Surfaces the most valuable next step.

Assessment-backed mastery.

Validates learning with real checks.