Skip to main content

Product Architecture

Myrm ships as two product layers plus optional cloud control plane. Understanding the split helps you deploy, debug, and migrate from other agents.

Layers

LayerRepo / packageLicenseWhat you run
Product UI & APImyrm-agent-frontend + myrm-agent-serverOpen sourceBrowser WebUI, REST/SSE API, auth, billing hooks
Agent harnessmyrm-agent-harnessProprietary runtime (bundled)Tool execution, memory, browser, sub-agents, compression
Control plane (optional)myrm-control-planeSaaS / self-hostLLM relay, Work Units, unified tool gateway
Desktop shell (optional)myrm-agent-desktopOpen source (Tauri)Native app wrapping the WebUI + local backend
Plain language: the frontend is what you click; the server is the door into your workspace; the harness is the brain and hands inside the sandbox.

Request flow (typical chat)

  1. User sends a message in frontend (or an IM channel via server).
  2. Server authenticates, loads agent profile, streams events to the UI.
  3. Harness runs the agent loop: tools, memory recall, sub-agents, context compression.
  4. Results stream back through server → UI (and optional channel notification).

What to configure where

TaskWhere
Models, API keys, agent personalityFrontend Settings → persisted by server
Skills, MCP, cron, goalsFrontend UI + server APIs
Memory browse / approve / deleteFrontend Memory panel
Sandbox files & code executionHarness inside per-user sandbox (not raw SSH)
SaaS billing & gateway toolsControl plane (when DEPLOY_MODE is SaaS)

Deployment modes (same features, different packaging)

ModeBest forYou get
Local WebUIDevelopers, BYOKFull GUI on localhost, data stays on your machine
Tauri desktopDaily driver on Mac/Win/LinuxApp icon in dock, deep links, no browser tab clutter
SaaSTeams without ops burdenManaged sandboxes, Work Units, optional tool gateway
PWAMobile add-to-homeInstall from browser after building myrm-agent-frontend (see Development Setup)

Migrating from other agents

  • Config: import paths for Hermes / OpenClaw-style exports (see Quick Start).
  • Memory: Myrm uses structured DB memory + GUI — not a single MEMORY.md file.
  • Skills: prebuilt + community discovery; evolution is approval-gated.
  • Channels: 25+ built-in providers — re-link OAuth/ tokens in Settings.
For an honest feature matrix vs OpenClaw, Hermes, Claude Code, and others, see Competitor Comparison.

Next

Quick Start

Run locally in minutes.

Memory System

How cross-session memory works.

Sandbox Runtime

Where tools actually execute.

Desktop App

Tauri packaging and updates.