Skip to content

Connect with Google

Add the MCP server, sign in with Google — no key to paste.

claude mcp add --transport http poremu https://api.poremu.ai/mcp
Build with MCP

Developer docs

Introduction

Poremu is an AI marketing platform with an API — products, strategy, calendar, contents, approvals, reports, automations, plus image, video and chat models. One key, one balance, and an MCP server for your assistant.

Poremu runs your marketing: you register a product, the AI marketer writes the strategy, fills the calendar, drafts contents, asks for approval before anything external changes, and sends reports. Everything the web app does is also an API — and every API call is also an MCP tool, so Claude Code, Claude, Cursor or ChatGPT can do the same work for you.

You want to…Use
Let your AI assistant run marketing (no code)Build with MCP — one URL, sign in with Google
Build products, strategy, calendar, contents, approvals and reports into your own systemMarketing API
Plug knowledge, links, accounts and data sources into a productResources
Run things on a schedule — weekly report by email, spend-spike alerts, daily AI check-upsAutomations
Get notified when something happens, or push events inWebhooks
Generate images and videos, or talk to an LLMImages · Videos · Chat · Apps

Base URL

text
https://api.poremu.ai

All endpoints are under /v1. Requests and responses are JSON (Content-Type: application/json) unless noted. Timestamps are ISO 8601 in UTC; dates (2026-11-05) are plain days.

Quickstart

  1. Create an API key in Settings → Developer → API keys (pk_live_…). It is shown once — store it as a secret. (Assistants connected over MCP do not need a key.)
  2. Register a product — paste a URL and Poremu reads the page, store listings or GitHub repo:
bash
curl https://api.poremu.ai/v1/products \
  -H "Authorization: Bearer $POREMU_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "workspaceId": "my-workspace",
    "name": "Lumo",
    "kind": "APP",
    "url": "https://lumo.app",
    "goals": ["INSTALLS", "SIGNUPS"],
    "monthlyBudgetKrw": 2000000
  }'
  1. Let the AI analyse it and write the seven strategy sections (positioning, messaging, audience, channel mix, KPI, budget, plan):
bash
curl -X POST https://api.poremu.ai/v1/products/PRODUCT_ID/analyze \
  -H "Authorization: Bearer $POREMU_API_KEY" -H "Content-Type: application/json" -d '{}'
  1. Fill the calendar from the recommended strategy template and read what it planned:
bash
curl -X POST https://api.poremu.ai/v1/calendar/fill \
  -H "Authorization: Bearer $POREMU_API_KEY" -H "Content-Type: application/json" \
  -d '{"workspaceId":"my-workspace","productId":"PRODUCT_ID"}'
  1. Ask the AI marketer anything — it reads metrics, drafts copy and files approval requests instead of changing budgets on its own:
bash
curl -N https://api.poremu.ai/v1/agent/chat \
  -H "Authorization: Bearer $POREMU_API_KEY" -H "Accept: text/event-stream" \
  -H "Content-Type: application/json" \
  -d '{"workspaceId":"my-workspace","productId":"PRODUCT_ID","message":"Compare the last 7 days and propose what to change."}'

The full route list is in Marketing API.

How it works

  • A workspace owns products, members and credits. workspaceId accepts the id or the slug; most people have one personal workspace, which is the default.
  • A product is the thing you market: brand kit, audiences, strategy docs (versioned per section), funnel, channels, campaigns, contents, resources.
  • External changes never run on their own. Budget changes, publishing, sending — whether triggered by the API, the AI marketer or an automation — become an approval that a workspace admin decides in the inbox (or an auto-approve rule). The decision executes through the connected channel.
  • Credits are the one currency: 1 credit = ₩10. Media is charged at the model's price, LLM chat per token, and the marketing engine adds small platform credits (analysis 1, content 1, report 5). API calls are free up to 1,000 per day. See Credits & pricing.
  • Long jobs — product analysis, agent runs, generations, app runs, automation runs — move through QUEUED → PROCESSING → SUCCEEDED | FAILED | CANCELED. Poll, long-poll with /wait, or subscribe with a webhook.

SDKs

There is no SDK to install — the API is plain HTTPS and works with any language. Examples use curl, JavaScript (fetch) and Python (requests). If you use an AI coding assistant, connect the Poremu docs MCP and it will write the integration for you — or connect the Poremu MCP itself and let the assistant do the marketing directly.