Make
Connect your apps step by step on a visual canvas and add AI where a decision is needed. Hosted only, with a free plan that never expires.
Make is a workflow tool with ai agent steps from Make (part of Celonis). It has a free option, paid use starts at $16/mo, it is not open source, and setup is medium. Availability: Most countries. We haven't tested it ourselves yet.
What it is
Make, formerly Integromat, is a hosted workflow tool owned by Celonis. You lay out a job as a "scenario" of steps on a canvas, like "new email arrives, read the attachment, add a row to a sheet". AI steps can read text, sort things and fill in fields, and Make AI Agents, still in beta, can decide which steps to call. Maia, a chat assistant inside Make, can draft a scenario from a plain description. There is nothing to install, and you pay in credits: most steps use 1 credit each time they run.
What it costs
Make plans
Sign up- Free$01,000 credits a month, 2 active scenarios, runs every 15 minutes at most
- Core$16/mo$12/mo billed yearly. 10,000 credits a month, unlimited scenarios
- Pro$28/mo$21/mo billed yearly. 10,000 credits, faster queue, full log search
- Teams$51/mo$38/mo billed yearly. 10,000 credits, team roles and shared templates
The Free plan has no time limit: 1,000 credits a month, 2 active scenarios, and scheduled runs at most every 15 minutes. Paid plans start at 10,000 credits a month, and the price goes up as you slide to more credits.
Affiliate link: if you sign up through it, we may earn a commission at no extra cost to you. It never changes our picks.
Checked Oct 20, 2026. Prices in US dollars before tax.
Should you pick it?
- Pick it if you want a job to run the same way every time, with AI only for the steps that need judgment
- Pick it if you don't want to install or look after anything yourself
- Pick it if you want to try a couple of small jobs on a free plan before you pay
Watch out for
- The Free plan allows only 2 active scenarios, runs no more often than every 15 minutes, and stops a run after 5 minutes.
- Every step uses credits, and AI steps can use more than 1. A busy scenario can burn through 10,000 credits fast, and when credits run out your scenarios stop until you buy more.
- Unused credits don't roll over. They expire at the end of the month, or the year on yearly billing.
- There is no self-hosted version, so your data always passes through Make's servers.
- Prices moved during October 2026. Check the credit slider on Make's pricing page before you pay.
Real use cases in our feed (2)
Ask an AI questions in Telegram by typing or sending a voice note, and get a short reply in seconds
A Make Community member wanted quick AI answers without leaving Telegram, so he built a Make scenario behind a Telegram bot. Each new message goes through a router: typed text goes straight to the AI step, while a voice note is first fetched from Telegram in two steps (get the file path, then download the audio) and turned into text. Both end up in the same AI step, with a short system prompt so replies stay easy to read on a phone, and the answer goes back to the same chat. He says it has run since August, handles around 50 messages a day, and replies within a few seconds. His tips: the two-step voice download is easy to miss, and filters on the router made it much easier to debug.
Check each supplier invoice against its purchase order, line by line, before anyone approves it
A poster in the Make Community built a scenario that picks up supplier invoices from Gmail, saves the PDF to Google Drive and reads the text with Google's OCR. Plain text rules pull out the invoice number and the purchase order number, and Make finds the matching order in a Drive folder. A Make AI Agent then compares the two line by line and marks each line as a match, a match with different wording, a mismatch or missing, with an overall status of approved, discrepancy found or escalated. The result and the agent's reasoning go into a Google Sheet. He tested it on three invoices: one matched despite different wording, one was flagged for a wrong quantity, and one was escalated because a line had no match.
Our test results
Not tested yet. We haven't run our own tests on this yet. When we do, results will appear here with the date, agent version, time, cost, how often we had to step in, and screenshots. How we test.