Agentic OS

Active skills

—

Brand context

Outputs today

Total sessions

Skills learned

Warnings

Pulse Log

—Awaiting heartbeat…

Brand Intelligence

Your core context — automatically injected into every skill execution.

Brand Voice

—

Ideal Customer

—

Positioning

—

Products & Services

—

Assets & Links

—

Start Here

Ten questions. Five minutes. Builds the intelligence layer that every skill reads before producing output.

What is your business and what does it do?

Name and one precise sentence — this becomes the identity anchor for every output.

1 / 10

Who is your ideal customer?

Role, industry, company size, situation. The sharper this is, the more precisely every skill will write for them.

2 / 10

What problem do you solve that no one else does quite like you?

Your real point of difference — honest answer, not the polished version.

3 / 10

How do you sound? Name 3–5 words that define your brand voice.

Every word choice in every output will reflect these traits.

4 / 10

What does your brand never sound like?

Tones, styles, phrases you actively refuse.

5 / 10

What is your tagline or core positioning sentence?

The sentence you use when someone asks what you do. Work in progress is fine.

6 / 10

What are your main products or services?

List each with a brief description — skills use these to frame your offers.

7 / 10

What has worked well in your marketing so far?

Formats, channels, specific messages. What has your audience actually responded to?

8 / 10

Paste a writing sample you love.

Your own best work, a competitor you admire, or any content that captures the voice you want.

9 / 10

Website, social handles, and visual brand references?

URLs, handles, brand colours, fonts — anything that anchors your identity.

10 / 10

Skill Registry

Copywriting

Generates brand-aligned copy in any format. Produces two complete versions so you can choose the best angle.

Configure Run

Format *

Goal *

Specific angle (optional)

Output as:

Output

Awaiting execution…

UGC Scripts

Writes high-converting short-form video scripts with production notes. Three complete scripts per run.

Configure Run

Script Type *

Platform

Length

Product to feature (optional)

Output as:

Output

Awaiting execution…

Content Repurpose

One source piece becomes many channel-ready formats. Paste any long-form content and select your outputs.

Source Content

Paste source content *

Formats to generate

Output as:

Output

Awaiting execution…

Trend Research

Identifies what is gaining traction in your niche. The brief feeds directly into your next copy or content run.

Configure Run

Time Frame

Platform Focus

Niche (optional)

Output as:

Research Brief

Awaiting execution…

Projects

0 outputs

▣

No outputs yet.
Execute a skill to generate your first project file.

Learning System

Per-skill feedback accumulated across sessions. Each skill reads its own section here before executing.

Memory

Session history. Short-term logs each session; long-term accumulates strategic knowledge.

Session Wrap-Up

Rate outputs, log feedback, and close the session. The system updates skill learnings and saves memory automatically.

Deliverables this session

No outputs generated yet this session.

Session notes

What worked well?

What to improve?

Next session priorities

Add Skill

Skill Name *

Category *

Purpose

Dependencies (comma-separated)

Reads Learnings?

Skill Steps

The Agentic OS System Blueprint

A complete reference for understanding, deploying, and scaling your AI-powered business operating system — from first principles to advanced workflows.

What is the Agentic OS?

The Agentic OS is not a collection of AI prompts. It is an operating system — a structured, self-improving intelligence layer that sits between you and AI, making every output brand-specific, contextually aware, and progressively sharper over time.

Most people use AI by typing a request and getting generic output. The Agentic OS inverts this. You build your intelligence once — brand voice, ideal customer, positioning — and every skill reads that foundation automatically before producing anything.

Core insight: the system is not business-specific. Only the brand context folder changes between businesses. The skills, the engine, the learning loop — all identical. Clone it for a nurse education company, a SaaS startup, or an e-commerce brand. Swap the context. Keep everything else.

The Three Layers

◈

Layer 1 — Brand Context

Built once during setup. Contains voice profile, ideal customer, positioning, products, assets, and writing samples. Every skill reads from here before producing output.

⌘

Layer 2 — Skills

Modular, category-organised, plug-and-play capabilities. Each is a SKILL.md with step-by-step instructions, reference files, and declared dependencies.

◑

Layer 3 — Learning System

SOUL.md defines agent identity. USER.md tracks preferences. learnings.md accumulates per-skill feedback. Memory stores session history. Everything compounds.

◉

The Heartbeat

thestoryofeveryheartbeat runs on every session start — scans skills, validates dependencies, syncs the registry, loads memory, and reports system status automatically.

How a Session Works

Run the Heartbeat

Opens every session. Scans skills, validates registry, loads context and learnings into working memory, reports status. Keeps everything in sync automatically.

Execute Skills

Choose a skill and run it. The executor automatically injects brand context, ICP, positioning, and the skill's own learnings before sending to AI. The output is already yours.

Save to Projects

Every execution saves to your projects library with timestamp, skill name, and full content. Review, copy, or open any output — nothing disappears between sessions.

Wrap Up

Ends every session. Rate outputs, log what worked and what to improve. The system writes those notes into each skill's learnings section. The next run reads them. This is the loop that compounds.

Use Cases by Business Type

🚀 SaaS & Tech

Product launch copy
Feature announcements
Cold outreach sequences
Case study repurposing
Onboarding email flows

🎯 Coaching & Consulting

Discovery call scripts
Programme sales pages
Newsletter content
Podcast repurposing
LinkedIn content system

🛍 E-Commerce & DTC

Product descriptions
UGC scripts for creators
Email abandonment flows
Ad copy variants
Influencer briefs

🏢 Agencies

Client proposal copy
Case study writing
Outreach email variants
Trend research reports
Social proof content

🏥 Healthcare & Education

Patient communications
Educational content
Referral letter templates
Course promo copy
Social media content

🎬 Media & Creators

Video script writing
Multi-platform repurposing
Sponsorship pitches
Trend-surfing content
Community newsletter

After the System Generates Output

After Copywriting — A/B test across channels

Run Version A in one campaign, Version B in another. Feed results back via Wrap Up so the system learns which angles work for your audience. Over 4–6 sessions, the skill converges on your highest-performing style.

After UGC Scripts — send to creators or film yourself

Forward scripts directly to UGC creators with production notes already included. Or use as teleprompter-ready scripts. The HOOK sections work as standalone ad hooks.

After Repurpose — build a full content calendar in minutes

One run gives you: newsletter, LinkedIn post, Twitter thread, 5 hooks, 5 email subjects, and 3 quote cards. That is a full week of multi-channel content from a single source piece.

After Trend Research — chain to execution skills

Run Trend Research then immediately run Copywriting or Repurpose — the system context now includes the trend data. Your next copy run incorporates trending angles automatically.

After multiple Wrap Ups — export Learnings as a strategy document

After 20+ sessions, your learnings.md is a living document of what works for your specific business. Use it as a brief for a copywriter, onboarding doc for a new team member, or quarterly strategy review.

Advanced Chained Workflows

Workflow	Chain	Outcome
Weekly content engine	Trend Research → Repurpose → Copywriting	Full week of multi-channel content in one session with current angles.
Launch sequence	Copywriting (awareness) → Copywriting (conversion) → UGC Scripts	Three-phase launch content — awareness, conversion, creator assets — all in brand voice.
Podcast monetisation	Repurpose (transcript) → UGC Scripts	One episode becomes a newsletter, LinkedIn post, Twitter thread, and three UGC scripts.
Competitor response	Trend Research → Copywriting (myth-bust)	Research competitor narratives then write counter-positioning copy that addresses the same pain from your angle.

Frequently Asked Questions

Everything you need to know about the Agentic OS — from getting started to advanced usage, troubleshooting, and scaling across multiple businesses.

Getting Started

What is the Agentic OS and how is it different from just using ChatGPT? +

The Agentic OS is a structured system, not a chat interface. The key differences are:

Persistent brand context — your voice, ICP, and positioning are injected automatically into every run. You never re-explain who you are.
Learning loop — the system logs feedback after every session and reads it before the next run. It gets smarter over time.
Skill registry — each task is a structured process (a SKILL.md file), not a freeform prompt. This means reproducible, consistently on-brand outputs.

With ChatGPT, every conversation starts from zero. With the Agentic OS, every run starts from everything you've already told it.

How long does the initial setup take? +

The Setup Wizard takes 5–10 minutes to complete. You answer 10 questions about your business — your voice, your customer, your positioning, your products, what's worked, and a writing sample. The system uses these to build your brand context files, which power every skill from that point forward. You only do this once per business.

Do I need any technical knowledge to use this? +

No. The webapp handles everything through the interface. You don't write code, edit files, or configure anything manually. The only technical requirement is an API connection for the AI execution — the rest is point-and-click. The Setup Wizard, all skill executions, the Wrap Up, and the learning system all work through the UI.

What is the Heartbeat and why do I run it at the start of every session? +

The Heartbeat (thestoryofeveryheartbeat) is the system's synchronisation engine. Every session it:

Scans the skill registry and detects new or removed skills
Validates dependencies for all active skills
Loads your brand context into working memory
Loads the learnings from previous sessions
Reports system status

Running it takes seconds and ensures the system is fully primed before any skill runs. Always run it first — it's what makes the outputs consistent.

Skills & Execution

Why does every skill produce two versions instead of one? +

Two versions is a deliberate design decision, not a limitation. Version A and Version B use different frameworks, hooks, or angles. This does three things:

Gives you real A/B testing material — deploy both and measure which resonates
Prevents you from anchoring on a single approach
Surfaces angles you might not have considered

After running a skill, rate both versions in Wrap Up. Over sessions, the system learns which approach performs better for your audience and skews toward it.

What is the HTML / Live Preview output format and when should I use it? +

Text mode gives you the raw content in a copyable format — best for copy you'll paste into your own tools (email platforms, social schedulers, docs).

HTML mode generates a complete, formatted HTML document you can download and use as a template or embed directly in a CMS.

Live Preview mode renders the HTML output in a browser frame inside the app — so you can see exactly what the page looks like before you use it. Best for landing pages, newsletter templates, and any content where visual presentation matters.

Can I chain skills together in one session? +

Yes — and this is where the system becomes most powerful. For example:

Run Trend Research → immediately run Copywriting — the trending angles are now in context
Run Repurpose on a transcript → run UGC Scripts — the source content informs the scripts
Run Copywriting (awareness) → Copywriting (conversion) → UGC Scripts for a complete launch sequence

All outputs go to Projects automatically. Wrap Up at the end logs what you built across the whole session.

How do I add a custom skill for something not already in the system? +

Go to Tools → Add Skill. Enter a name, category, purpose, dependencies, and the skill steps. The system generates a SKILL.md preview you can review before registering. Once registered, the Heartbeat picks it up on the next session start — no other configuration needed. Custom skills support the same learning loop as built-in skills: they get a section in learnings.md and improve with every Wrap Up.

The Learning System

What exactly is the "learning loop" and how does it actually improve outputs? +

The learning loop works in four steps:

Run — skill executes with full brand context injected
Review — you rate the output and leave notes in Wrap Up
Log — notes are written to that skill's section in learnings.md
Apply — next time the skill runs, it reads its own learnings before executing

This means the skill reads things like "tone was too formal last session — make it more conversational" before it starts writing. After 5–10 sessions of consistent feedback, outputs require significantly less editing.

What happens if I don't complete the Wrap Up at the end of a session? +

Your outputs are still saved to Projects — that happens automatically. What you miss is:

The feedback being logged to learnings.md
The skill improvement that would have happened
The short-term memory entry for that session

The system won't break, but it won't learn from that session. If you're short on time, even a minimal Wrap Up — just a rating and one note — is better than none. The compounding only happens if you close the loop.

How many sessions before I notice a meaningful improvement in output quality? +

Most users notice a difference by sessions 3–5 if Wrap Up is completed consistently. By sessions 10–15, the system typically produces outputs that require minimal editing — it has converged on your specific voice, your audience's language, and the angles that work for your business. The rate of improvement depends directly on how specific your feedback notes are. "Make it shorter" is less useful than "Cut the second paragraph — the CTA is strong enough on its own."

Business & Scaling

Can I use this system for multiple businesses or clients? +

Yes — the system is designed for this. The skills, engine, and learning infrastructure are identical across businesses. Only the brand context folder changes. To deploy for a new business or client:

Run the Setup Wizard again (it overwrites brand context only)
Or keep separate browser profiles / localStorage instances for each client
Skills persist and are shared — new business just gets different brand context injected

This makes the Agentic OS viable as an agency operating system, with one skill set serving multiple client contexts.

How do I know if my brand context is set up correctly? +

Run the Copywriting skill on a Social Post and compare the output to something you'd actually publish. Ask:

Does it sound like you, not like generic AI?
Does it speak to your actual customer, not a hypothetical one?
Would you edit out more than 20% of it?

If the answers are no, yes, and no — your brand context is working. If outputs feel generic, the most impactful fix is improving your writing sample (question 9 in the wizard). The AI calibrates from real examples more accurately than from descriptive adjectives.

What's the difference between Projects, Learnings, and Memory? +

Projects — the archive of every output generated. Raw content, searchable, with full text on click. This is your content library.

Learnings — per-skill feedback logs. What worked, what to improve, your ratings from each session. This is what makes skills smarter over time.

Memory — session-level records. What you worked on, decisions made, next priorities. Short-term tracks recent sessions; long-term accumulates strategic knowledge that doesn't expire. Think of Projects as your content archive, Learnings as your training data, and Memory as your notes between sessions.

Is IamWealthy a real person or an AI? +

IamWealthy is your AI business advisor — a specialist chatbot built into the Agentic OS. It is powered by the same Claude AI that runs all the skills, but with a specific persona focused on business growth, strategy, and wealth-building. You can ask IamWealthy about how to use any skill, what workflow to run for your situation, how to interpret your outputs, or any business question. Find it in the chat bubble at the bottom-right corner of the screen.

Output Formats

When should I use HTML output vs. plain text? +

Use plain text when: you're copying output into a tool that handles its own formatting (email platforms, Notion, social media schedulers, Google Docs).

Use HTML when: you need a formatted document to embed in a CMS, send as a branded email, or use as a template for a developer.

Use Live Preview when: you want to see exactly how the content looks rendered before using it — ideal for landing page copy, newsletter templates, and any content where visual hierarchy matters. You can copy the HTML from the code view and paste it directly into any HTML-capable platform.

Can I copy the generated HTML and use it on a real website? +

Yes. The HTML output is complete, self-contained, and production-ready. It includes inline styles (no external dependencies), so it works anywhere HTML is accepted — email clients, CMS pages, landing page builders, and raw HTML files. Switch to HTML mode, copy the output, and paste it wherever you need it. If you want to customise colours or fonts, the HTML uses CSS variables at the top you can easily edit.