Welcome to the Moonage documentation. Learn how to set up and use the AI operations layer for your team.

Welcome to Moonage

Moonage is the AI operations layer that connects your tools and automates workflows using intelligent agents. Each agent lives in a Space — a persistent environment with its own context, integrations, and personality.

Why Moonage?

Modern teams use dozens of tools — Notion for docs, Linear for issues, Slack for communication, GitHub for code. Context is scattered. Moonage brings it all together into a single, intelligent layer that understands your data and can act on it.

How it works

Your Tools → Moonage Agent → Automated Workflows
  Notion        Space          Smart Actions
  Linear        Context        Schedules
  Slack         Persona        Event Triggers
  GitHub        Memory         Parallel Execution

The agent indexes your connected tools, builds a context graph of entities and relationships, and uses that understanding to execute multi-step workflows — either on demand, on a schedule, or triggered by events.

Core capabilities

Connect your tools

OAuth integrations with Notion, Linear, Google Drive, Slack, GitHub, and more. No API keys needed — connect in seconds, and the agent starts indexing immediately.

Create Smart Actions

Multi-step automations that run across integrations. Ask the agent to "summarize all open issues in Linear and post a digest to Slack" — it handles the rest, executing independent steps in parallel.

Schedule workflows

Set up recurring automations with a visual schedule builder. Weekly standups, daily digests, monthly reports — all automated.

Ask questions

Your agent understands your data in context. Ask "What did the team discuss about the Q1 roadmap in Slack last week?" and get answers that traverse connections across tools.

Execute code

Moonage can run code in a sandboxed environment. Generate charts, transform data, run calculations — all within a secure Cloudflare sandbox.

Architecture overview

Moonage is built on Cloudflare Workers for global low-latency execution:

Component	Technology	Purpose
Orchestrator	Workers + Durable Objects	MCP protocol, session management
Agent	Workers	LLM routing, tool orchestration
Brain / Retrieval	Workers + R2 + SuperMemory	Scoped memory, semantic retrieval, and file intelligence
Auth	Workers + KV	OAuth management, token refresh
Content	Workers + D1 + R2	Blog, changelog, docs

Request flow

Intent analysis — classify the query and identify relevant tools (Haiku, ~5s)
Catalog discovery — find available tools from connected integrations (KV-cached)
Execution planning — LLM produces a dependency DAG of tool calls (Sonnet)
DAG execution — run tool calls in parallel waves, re-plan after each wave
Result collection — package raw MCP data as structured payload

Quick links

Getting Started — Set up your first Space in under 5 minutes
Core Concepts — Understand Spaces, Smart Actions, and the context graph
Integrations — Connect your tools
API Reference — Build on top of Moonage programmatically

Introduction