What Make is
Make is an automation platform that connects web apps without code. It began as Integromat, a startup from the Czech Republic. Celonis, a process-mining company, bought it in October 2020, and Wikipedia says the product has since been renamed Make. Wikipedia files it under iPaaS, an integration platform delivered as a service you reach in a browser.
You build a workflow, called a scenario, on a canvas. Each step is a module: a trigger that starts the scenario, then modules that read data, change it or write it to another app. Lines join the modules, and filters and routers split the flow. A February 2026 launch video from Make describes its AI Agents as built inside the same canvas and working across more than 3,000 apps. A free plan exists, and paid plans raise the monthly credits.
What Make is good for
- Multi-step flows you can see: the canvas shows every module and the lines between them, so you can follow where a value goes.
- Branches: a filter lets only matching items continue, and a router sends one item down several paths, as the beginner tutorial from Metics Media shows.
- Sorting messages with AI: Make's own first tutorial watches Gmail for support emails, asks ChatGPT to file each as Billing, Account Management, Legal or Other, and logs it in Google Sheets.
- Instant or timed starts: a scenario can start from a webhook, from a change in an app, or on a schedule you set.
- AI agents on the same canvas: a 2026 course walks through giving an agent a model, instructions, tools and memory.
- Letting an AI assistant run your scenarios: Make's MCP server exposes scenarios set to run on demand as tools that a client such as Claude Desktop can call.
When to use Make, and when to pick another tool
| Your job | Use Make? | Or try |
|---|---|---|
| See a flow with branches drawn out | Yes. The canvas and its routers are its main idea | Zapier, which lists steps one under another |
| Connect two apps in the quickest way | Yes, and so does the other hosted tool. Compare the free plans first | Zapier, which counts tasks and has a free plan |
| Keep workflow data on your own server | No. Make runs as a hosted service | n8n, which you can self-host for free |
| A flow that runs thousands of times a day | Check the credit count first, since the free plan includes 1,000 a month | n8n, which costs nothing to run yourself |
| Let an AI assistant call your flows as tools | Yes, through its MCP server | MCP servers built for one app |
| Chat, write or research | No | ChatGPT, Claude or Perplexity |
| A custom app with its own code | No, Make connects apps and does not replace them | Cursor or Claude Code |
How to use Make, step by step
These steps are for Make's website. Menus move when the app updates, so they stick to the parts that stay put: the scenario, its modules, a test run and switching it on.
- Create a free account on Make's website. Its pricing page, as read for the agent builder lesson, lists a free plan with 1,000 credits a month, so check which plan your account is on.
- Open the Scenarios page and start a new scenario. It opens as an empty canvas with a place to add the first module.
- Add the trigger module. Search for the app, such as Gmail, pick the event that should start the scenario, such as a new email, and connect your account when Make asks.
- Press Run once and cause one real event, such as sending yourself an email. Make reads it, and the trigger now holds real values for the next steps to use.
- Add a second module. For an AI step, search for an OpenAI or ChatGPT module and fill its prompt by clicking fields from the trigger in the mapping panel, not by typing them.
- Add an action module, such as Google Sheets and a new row. Map the trigger fields and the AI answer into the columns. To stop some items, select the line between two modules and add a filter.
- Run once again. Open the bubble on each module to read what went in and what came out, and fix any module that shows an error.
- Set the schedule, then switch the scenario on with its toggle. After the first real event, open its execution history to confirm it ran and to see the credits it used.
Open the Scenarios page and press Create a new scenario.
The new scenario opens as an empty canvas. Press the big + in the middle to add the trigger module, such as Gmail and a new email.
With the modules in a chain, press Run once and cause one real event, such as sending yourself an email.
Set the schedule, then switch Scheduling on so the scenario runs by itself.
AI-made illustrations of Make's screens, drawn on 6 November 2026 to match these steps. They are not screenshots, so the real app may look different: follow what you see.
The tutorials at the top of this page show these steps on screen. Make changes its menus now and then, so if a name differs, follow the one you see.
An example to try
Put this prompt in the AI module of a scenario that watches a mailbox. Replace the two bracketed lines by clicking the matching fields from the Gmail module. Use a spare mailbox, since the scenario acts on real mail once you switch it on.
Read this support email and answer with one word only: Billing, Account Management, Legal or Other.
Subject: [subject from the Gmail module]
Text: [text from the Gmail module]- Make's own tutorial builds this idea with Gmail, ChatGPT and Google Sheets, and it runs only during office hours.
- The answer is one word, so a later filter or router can compare it to a fixed list without reading a sentence.
- Nothing runs on its own until you switch the scenario on, so test with Run once first.
Tips for better results
- Plan the flow on paper first. List the trigger, each action and what happens to items that do not match, then build it.
- Start with a trigger and one action. Add filters, routers and an AI step after the plain version works.
- Name each scenario for what it does, so the list stays readable when you have ten.
- Use a filter when you only want to skip some items, and a router when each kind of item needs its own action.
- Check the credits after a busy week. A scenario that starts on every new row or message can use up a monthly allowance early.
Limits and costs
- Make counts usage in credits. Its pricing page, as read for the agent builder lesson on 6 November 2026, lists 1,000 credits a month on the free plan and says some AI features use more than 1 credit.
- The same page lists AI Agents on every paid plan, so check which plan holds the features you need before you build on them.
- Make is a hosted service, and that page does not describe a self-hosted version. For data that must stay on your server, n8n fits better.
- A scenario that is switched on acts on its own, so test with spare data before you connect a real customer mailbox.
- The 3,000 app figure comes from a Make video description in February 2026. If an app is missing, check its page on Make before you plan around it.
Learn more about Make
- 6Best AI agent builder in 2026: ready-made agents, no-code builders and SDKs comparedCompare ready-made AI agents and agent builders: ChatGPT Work, Claude Cowork, Gemini, Manus, n8n, Zapier, Make, Copilot Studio and SDKs on price and limits.
- 1n8n tutorial for beginners: build your first workflowAn n8n tutorial for beginners: build a first workflow with a schedule trigger, an HTTP Request, an Edit Fields node and a data table, then test and publish it.
Frequently asked questions
Is Make free?
Yes, it has a free plan. Make's pricing page, as read for the agent builder lesson on 6 November 2026, lists 1,000 credits a month on it. Paid plans raise the credits, and some AI features use more than 1 credit.
What is a scenario?
A scenario is Make's name for one automated workflow. It starts with a trigger, then runs modules in order, and filters or routers can split the flow along the way.
Is Make the same as Integromat?
Yes. Wikipedia says Celonis bought the Czech startup Integromat in October 2020 and that the platform has since been renamed Make. Older tutorials and forum posts may still call it Integromat.
Do I need to code to use Make?
No. You pick apps, set modules and map fields on a canvas. Code only matters if you want a custom step, and the plain version of most flows needs none.
Should I start with Make, Zapier or n8n?
Start with Make if you want a visual canvas in the cloud with a free plan. Pick Zapier for a step list with its own free plan, and n8n if you want to host it yourself for free. The agent builder lesson compares all three.
Can ChatGPT or Claude use Make?
Yes, through MCP. Make's GitHub page for its MCP server says it turns scenarios set to run on demand into tools that AI assistants can call, and a newer cloud version is linked from that page. A separate Make repository on GitHub offers skills for Claude Code, Codex and Cursor.
Sources
- Wikipedia: Celonis https://en.wikipedia.org/wiki/Celonis
- Make: Pricing https://www.make.com/en/pricing
- Make: Make vs Zapier https://www.make.com/compare/make-vs-zapier
- GitHub: integromat/make-mcp-server https://github.com/integromat/make-mcp-server
- GitHub: integromat/make-skills https://github.com/integromat/make-skills
- Make on YouTube: New Make AI Agents launch https://www.youtube.com/watch?v=4JsSl6JjmAs
- Make on YouTube: Create Your First Scenario in Make https://www.youtube.com/watch?v=oigckzWPY60
- Metics Media on YouTube: Make.com Tutorial for Beginners (2026) https://www.youtube.com/watch?v=SVkiqiSVo3k
- Emmy White Experience on YouTube: Make.com Tutorial for Beginners https://www.youtube.com/watch?v=zns3VuyjbW8
- Dimitris Solutions on YouTube: How To Build AI Agents In Make.com https://www.youtube.com/watch?v=0h9H04Ar_C4
- Zapier: Pricing https://zapier.com/pricing
- n8n: Pricing https://n8n.io/pricing/