In this article I will show you my openclaw agent for creem(MoR) to perform & monitor various operations of creem store, without checking the dashboard.
But before we jump into the setup, let’s quickly understand what OpenClaw actually is.
What is OpenClaw?
OpenClaw is an open-source AI agent framework that lets you build your own custom agents to automate workflows, connect tools, and interact with external services and most interesting thing is you can directly control the agent from the telegram/whatapp/slack.
We will create our agent with openclaw that continuously monitor our creem store and gives us alerts whenever required.
Here’s what we expect from the agent:
- New payments and failed payments alerts
- Churn detection alerts.
- Daily digest that gives overall overview of the store.
- Heartbeat monitoring on every 4-5 hours according to the state of store.
- Answer any the question you have regarding your store.
Installation
Install openclaw on your local system. visit the official site for more details: https://openclaw.ai/
I have connected it with my telegram bot and I am using OpenAI for interaction with the agent.
You also need to install the Creem CLI and need to login with your api key. for instructions visit: https://docs.creem.io/code/cli
Workflow 1: Real time alerts like new payment, failed payment and churn detection events.
To give realtime alerts it is mandatory that we receive the webhooks from the creem. Openclaw itself can’t receive the webhooks. So, we have created a small nextjs application that has only route /api/webhooks/creem . It just receives the webhook and makes a post request to our openclaw hook. so from here our openclaw receives the message analyse it and gives alert to the telegram. You can check the code of this nextjs application here: https://github.com/DeepKumbhare/creem-openclaw-demo-app/tree/main
So for setup, you need to clone this repo and start dev server.For receiving the webhooks, you need to use ngrok.
Now, when I will make the payment in a product, I will receive the alert like below in my telegram bot:
I will also receive the similar alerts when payment fails or when we detect events like subscription cancelled or expired(churn detection alerts). I am attaching the screenshot below for more details:
Workflow 2: Heartbeat monitoring on every 4-5 hours according to the state of store.
I have setup the monitoring of store using the heartbeat. Here is the official file creem provides for heartbeat: https://creem.io/HEARTBEAT.md
It saves the state everytime it runs locally and gives us alert about the changes between two runs. We can define the time period we want to run the heartbeat. Ideally 4-5 hours is recommended.
Workflow 3: Daily digest of the store:
It gives us the overall summary of the store in the day. You can set like I want daily digest in morning 9 AM. So, Openclaw will give the daily digest everyday 9 AM in morning.
Here is how daily digest looks like:
This contains almost every aspect of our store like revenue, churn , subscription health, risks and suggested action.
Workflow 4: Ask anything about your store.
As we have given it our creem skill and it has access to our creem account via creem CLI, our agent can answer any of your questions related to your creem store
For example: You can ask it to list all the products of your store or you can ask to generate the discount code that you can share with your users.
I have also asked the similar questions. I am attaching the screenshots below:
Repositories:
Demo application for receiving the webhooks: https://github.com/DeepKumbhare/creem-openclaw-demo-app/tree/main
Agent configs: https://github.com/DeepKumbhare/agent-configs/
You can easily use my agent configs and setup similar agent for your creem store in openclaw.
Summary
The article describes an AI agent built with OpenClaw to monitor and manage a Creem (Merchant of Record) store automatically, without needing to check the dashboard manually.
OpenClaw is an open-source AI agent framework that connects to messaging platforms like Telegram, WhatsApp, or Slack, allowing users to control agents conversationally.
The agent handles four main workflows:
Real-time alerts — detects new payments, failed payments, and churn events (cancelled/expired subscriptions) sends notifications to users via Telegram.
Heartbeat monitoring — runs every 4–5 hours, compares the current store state to the previous one, and alerts the user about any changes.
Daily digest — sends a morning summary (at a configured time like 9 AM) covering revenue, churn, subscription health, risks, and recommended actions.
Q&A about the store — since the agent has access to the Creem CLI, users can ask it anything about their store, such as listing products or generating discount codes.
Top comments (0)