AIOHM Platform

Appearance

Muse Deep Dive ​

Muse is the central AI operating partner inside AIOHM. This page gives the production /docs/muse/ route a dedicated home and explains how Muse fits into the platform.

Overview & Two Modes ​

Muse works in two distinct modes:

  • Private platform mode for internal strategy, planning, creation, and operations
  • Public mirror mode for audience-facing questions that should only use approved public knowledge

That split lets one brand keep an internal thinking partner and a public support surface without mixing private context into public answers.

All 11 Agents ​

The platform combines Muse with specialist agents so work can stay focused when a job crosses domains.

  1. Muse
  2. Marketing Agent
  3. SEO Agent
  4. WordPress Agent
  5. WooCommerce Agent
  6. Mautic Agent
  7. Analytics Agent
  8. Web Builder Agent
  9. RSS Autoblog Agent
  10. Mirror Mode Assistant
  11. Entity Extraction Agent

Muse is the operator-facing hub; the others provide domain depth when specific tools or prompts are needed.

Three-Tier Memory ​

Muse keeps context through a layered memory model:

  • Immediate conversation context for the active session
  • Summarized session memory for continuity across longer work
  • Durable tenant knowledge from documents, notes, and brand data

This structure helps Muse stay useful over time without forcing every conversation to start from zero.

The Gardener ​

The Gardener is the maintenance mindset behind memory quality.

  • Keep useful facts alive
  • Compress repeated patterns into smaller summaries
  • Reduce clutter so high-value context stays available when it matters

In practice, this means memory should evolve with the brand instead of becoming a noisy archive.

Agent Skills ​

Skills let Muse and other agents inject reusable operating knowledge at the right time.

  • Skills can define playbooks, checklists, and standard procedures
  • They help shape responses without rewriting the full system prompt every time
  • Tenant-specific skills make each workspace behave more like a trained operator than a generic chatbot

Sub-Agent Delegation ​

Muse can delegate to specialist agents when a prompt needs deeper domain handling.

  • WordPress tasks can move into the WordPress agent and its tooling
  • SEO analysis can move into the SEO agent
  • Campaign or marketing tasks can move into focused specialist flows

This keeps Muse strategic while still giving users one main place to start work.

File Upload Context ​

Uploaded files can become part of the AI working context.

  • Documents can be scanned and summarized
  • Relevant excerpts can be reused inside conversations
  • File-driven prompts become much more precise than freeform chat alone

Python Sandbox ​

When enabled, the platform can expose a controlled Python execution path for tasks that benefit from structured computation.

  • Useful for data shaping and limited analysis work
  • Should stay tightly scoped and safety-checked
  • Complements the rest of the agent tooling instead of replacing it

Why Muse Matters ​

Muse is not just another chat box. It is the tenant-aware operating layer that connects memory, tools, documents, brand voice, and execution workflows into one system.