AI Marketing Agent

August 26, 2026 · View on GitHub

An AI agent for marketing strategy — email campaigns, multi-channel content planning, and product marketing — backed by real campaign and analytics APIs.

Part of Agency Agents OS, an open ecosystem of specialized AI agents for real business work.

Agency Agents OS — the central catalog this repo is part of.
ai-ads-agent — paid-media counterpart to this repo's organic marketing sub-agents.
ai-social-agent — publishes the campaigns this repo plans.
ai-content-repurposing-agent — turns campaign source video into supporting clips.
MuAPI MCP docs — connect this repo's SKILL.md files via MCP.
MuAPI Agent Skills — background on the SKILL.md pattern this repo uses.
MuAPI access keys — create the API key this agent needs.

What this covers

This repo is the generic marketing entry point of Agency Agents OS: lifecycle and campaign email drafting, multi-channel content campaign planning, and product positioning/launch-asset planning. It is not the place for paid media or organic social — those get their own dedicated umbrellas so each agent's workflow, connections, and approval boundaries stay scoped to one job.

Paid media (ad copy, targeting, spend) lives in ai-ads-agent.
Organic social (posting, scheduling, community) lives in ai-social-agent.

Sub-agents

Agent	What it does	Status
Email Marketing	Drafts and sequences lifecycle and campaign emails from a brief	Coming Soon
Content Campaign Strategy	Plans a multi-channel content campaign around a launch or theme	Coming Soon
Product Marketing	Positioning, messaging, and launch-asset planning for a new product or feature	Coming Soon

Required Muapi APIs

These sub-agents are planned against the following Muapi capability names. None are live yet — see Status below.

marketing.send_email — dispatch drafted email sequences through a connected sending capability
marketing.campaign_analytics — pull campaign performance (opens, clicks, conversions) to inform strategy
analytics.ga4_report — pull site/funnel analytics to ground content and positioning decisions in real traffic data

Setup

Create a Muapi account and API key at muapi.ai.
Review the full capability surface in the Muapi OpenAPI schema: https://api.muapi.ai/openapi.json
Load the SKILL.md file for the sub-agent you want into your agent runtime (Claude, or any other SKILL.md-compatible runtime).
Provide the required inputs listed in that sub-agent's SKILL.md (brief, audience, launch details, etc.).

Using with an AI agent

Every sub-agent's SKILL.md is model- and runtime-agnostic — it's plain Markdown, so it works with any LLM agent, not just Claude. Two integration paths:

As an MCP connection (the agent gets live Muapi tools):

Muapi runs an MCP server at https://api.muapi.ai/mcp that any MCP-compatible client can connect to — Cursor, Windsurf, Claude, or your own custom agent.

Cursor / Windsurf / other clients with a header field: connect to https://api.muapi.ai/mcp with an Authorization: Bearer YOUR_MUAPI_KEY header.
claude.ai / Claude Cowork / other connector UIs with no header field: use the URL-embedded key form instead, https://api.muapi.ai/mcp/YOUR_MUAPI_KEY, via Settings → Connectors → Add custom connector.
Claude Code / Claude Desktop: claude mcp add muapi -e MUAPI_API_KEY=YOUR_MUAPI_KEY -- muapi mcp serve (uses the muapi CLI's stdio transport — Claude Code's HTTP MCP client doesn't reliably inject tools).

Full setup details for every client: muapi.ai/docs/mcp.

As agent instructions (any LLM follows the workflow directly):

Drop a sub-agent's SKILL.md into a Claude Code project's .claude/skills/ directory, paste it into a custom-GPT/Project's system instructions, hand it to an autonomous agent framework as a tool spec, or attach it directly in a chat conversation — then ask the agent to follow it.

Read-only vs. write actions

Email drafting is draft-only. The agent produces subject lines, copy, and sequencing — it does not send anything.
Sending is requires-approval-to-publish. Dispatching a drafted sequence through marketing.send_email always requires an explicit human approval step before anything goes out.
Content and positioning planning are inherently draft/plan outputs — they produce documents for a human to review, not published artifacts.

Status and limitations

All three sub-agents are Coming Soon. They are blocked on marketing.send_email, marketing.campaign_analytics, and analytics.ga4_report — none of which are live on Muapi yet, since Muapi's current API surface is focused on generative media (image, video, audio) rather than marketing/analytics tooling. Once these capabilities ship, each SKILL.md will be updated from a planning document into a working agent definition.

Contributing

See the central CONTRIBUTING.md in Agency Agents OS for how to propose changes across the ecosystem.

License

MIT — see LICENSE.