🔁
The same check, every morning
You open the same screen, copy the same numbers, and share them in the same words.
Sign in with Google — free for individuals and businesses
Your work
Your work
has seams.
eeumsae holds
them together.
Call the APIs of the services you already use, keep only the values that matter, and let AI summarize them into KakaoTalk or Slack. No server to run, no cron job to babysit — switch it on once and it keeps going.
Sign in with Google at app.eeumsae.com and start right away. See the docs for how it works, or write to [email protected]. The app interface is currently in Korean.
이음새 · eeumsae
The seam where separate things meet.
이음새 (eeumsae) is Korean for the seam — the place where two pieces are joined. It takes over the work you have been doing by hand between one tool and the next: APIs, data, notifications, and AI, in one flow.
Why eeumsae
Does your day look like this?
🖥️
A server, just for this
Standing up a server, wiring a cron, building failure alerts — all to run one script. It should not take that.
🧩
AI starts over every time
It does the job well in the chat window, but it will not do it again tomorrow morning on its own.
How it works
Switch it on. eeumsae takes it from there.
There are two ways to build one: ask your AI assistant in plain words, or connect the nodes yourself in the editor.
-
01
Ask for it, or connect it yourself
Connect eeumsae to an AI assistant such as Claude, and “every day at 9, check this and send it to KakaoTalk” is enough to get a first draft.
-
02
Fill in connections and values
Connect Slack, and keep sensitive values like API keys in workspace variables — referenced, never pasted into the workflow.
-
03
It runs exactly as set
It runs on a schedule or a webhook. If it fails, it retries, then notifies the webhook you registered.
See before you run
Every automation, at a glance.
Review what the AI built before you switch it on — what runs, when it runs, and where it goes out.
What makes it different
Automation is only worth it if you can trust it.
⚡
If it has an API, it connects
Fill in a URL, headers, and a body, and any REST API becomes a node. Pick just the values you need out of the response and pass them on.
🛡️
Every time, the same way
It runs in the order you set and retries as many times as you allow. If it still fails, it does not pass quietly — it tells you.
✋
Approval before anything goes out
When AI wants to run an automation that sends something, it shows you what goes where and waits for your approval.
🤖
From your AI assistant
Connect it to an AI assistant such as Claude: say what you want, and eeumsae builds it and runs it.
For example
Only on days something changed,
this lands at 9 AM.
It checks the latest release of every library you watch, every day, and tells you only when something differs from last time. The same list never shows up twice.
- Calls the GitHub releases API at a set time, once per entry on the watchlist
- Compares each with the saved tag and keeps only what is new
- Sends to your connected channel only when there is a change, retrying as many times as you allow
release-watch
📦 2 releases changed
spring-boot v4.1.1 and vite [email protected] are out. The other 3 are unchanged since last time.
AI assistant · MCP · Skill
Nothing new to learn.
You do not have to learn eeumsae. Use the Claude subscription you already pay for, and build and run automations just by asking — the way you already do.
- The AI you already use — paste the eeumsae server address into a Claude custom connector (MCP). In Claude Code it is one command.
- Built and checked — it turns what you said into a workflow and validates it before saving.
- Run it and look back — hand over execution and review the history. If something gets sent, it asks first.
AI decides. eeumsae runs it.
Pricing
It costs nothing to use.
The cloud service is free, whether you are an individual or a business. If it has to run inside your own infrastructure, ask about the self-hosted edition.
Cloud
Free
Individuals and businesses alike
- Sign in with Google at app.eeumsae.com and start right away
- Workflows, runs, connections, and the AI assistant (MCP) — all included
- Fees for external APIs or LLMs are paid to those services directly
Self-hosted
Contact us
On-premises · Private cloud
- Installed on your own infrastructure (Kubernetes or Docker)
- For environments where data must not leave your network
- Installation and operations are worked out over email
You can start right now.
Sign in with Google and switch on your first automation. Free for individuals and businesses alike. Questions are always welcome.