Visual automation · checked 28 Sept 2026
DeskFerry vs Make
Make lays an automation out as a scenario: round modules on a canvas, joined by lines, split by routers and gated by filters. DeskFerry skips the diagram. You write down the job and an agent works out the route each time.
Choose Make when
- you think visually and want every branch of the logic on one screen
- you run high volumes of simple syncs, where Make's credits are cheap
- you want to choose which AI model runs each step, with your own key
Choose DeskFerry when
- the routing depends on reading the content, not matching a field
- nobody on the team wants to learn routers, iterators and filters
- you want the agent to ask a person before it posts or pays anything
01 · Setup
Make: draw a scenario of modules and routers
DeskFerry: write a brief, pick the apps
02 · Metered on
Make: credits: 1 per module run
DeskFerry: credits: per job, by effort
03 · App library
Make: 3,000+ apps
DeskFerry: 1,500+ apps
04 · Entry price
Make: free 1,000 credits, then from $9/mo
DeskFerry: 7-day trial, then $49/mo
Same job, built twice
Supplier invoices arrive by email. Check them against the PO and file them.
In Make this is a scenario with a router: one branch for invoices that match, one for the ones that don't. In DeskFerry the branching lives in the brief.
Map: supplier, PO number, total, due date
Every branch is visible, and a mismatch is caught only if you wrote the filter for it. A new invoice layout usually means adjusting the field mapping.
Brief
When a supplier invoice arrives, read it, find the matching PO in the tracker sheet and create the bill in Xero. If the total, supplier or PO doesn't line up, don't file it; tell #finance what's off.
Allowed apps
Last run
- Read INV-2291 from Harbour Print (2 pages, scanned)
- Found PO-0457 in the tracker: $4,180.00
- Invoice total is $4,318.00: a $138 freight line is not on the PO
- Held the bill and wrote up the difference for #finance
- Asked in #finance whether to approve the freight charge
The agent explains why it held an invoice instead of dropping it down a branch. You trade a fully drawn map for a written rule.
Feature by feature
Where each one wins
| Feature | Make | DeskFerry | Edge |
|---|---|---|---|
| How you build | Visual scenario canvas; Maia can draft one from a prompt (beta) | Plain-English brief | Depends |
| Seeing the whole logic | Every route and filter on one canvas | Written in the brief; each run is logged | Make |
| App library | 3,000+ apps | 1,500+ apps | Make |
| Choosing the AI model | 400+ AI app modules; bring your own key on paid plans | Managed for you | Make |
| Input that varies run to run | Needs a filter or route per case | The agent reads it and decides | DeskFerry |
| Agents | Make AI Agents, in open beta since Feb 2026 | The core product | DeskFerry |
| Users | Unlimited users per organisation | Per plan, not per seat | Depends |
| Free option | Free plan, 1,000 credits a month | 7-day free trial | Make |
Sources: make.com/pricing · make.com/integrations · Make help: credits · Make help: AI Agent · checked 28 Sept 2026
How the bill works
Both sell credits. They count different things.
On Make, one module run is one credit, so a seven-module scenario costs about seven credits every time it fires. AI modules are metered by tokens instead. At $9 for 5,000 credits, simple high-volume syncs are cheap.
On DeskFerry, credits are spent per job and depend on the work involved; a simple task takes one to five. You pay more per credit, and in return no one maintains the routes.
A worked example: Make is the cheaper line here, and that is the honest result for straightforward volume. Real use on both depends on the job.
Straight talk
When to stay on Make
Your ops lead lives on the canvas
If someone already reads Make scenarios fluently, the visual map is an asset, not overhead. Keep it.
High-volume, low-judgment syncs
Moving thousands of rows a month between two apps is what Make's per-module credits are priced for.
You want model control
Bring-your-own-key on paid plans and hundreds of AI modules give you finer control over which model runs where.
Moving over
Move the scenario with the most routes
Scenarios with three or more routes and a stack of filters are usually trying to encode judgment. Those are the ones worth moving.
Write down each route in words
One line per branch: when it fires and what should happen.
Merge the lines into one brief
The routes become sentences; the filters become the conditions the agent checks.
Switch the scenario off after a week
Run both, compare what each did with the edge cases, then retire the scenario.
FAQ
Make questions
Make has AI agents now. How are they different?
Make AI Agents have been in open beta since February 2026 and run on the same credits as scenarios. They are a new object inside Make's builder. DeskFerry is built only around agents, with briefs, approvals and Slack runs as the defaults.
Is Make cheaper than DeskFerry?
Per credit, yes: $9 buys 5,000 Make credits. But a Make credit is one module run, and a DeskFerry credit pays for part of a whole job. For simple, high-volume syncs Make usually costs less.
Did Make change from operations to credits?
Yes. On 27 August 2025 Make renamed operations to credits at a 1:1 rate for standard modules. AI modules moved to token-based pricing later that year.
Can I keep Make and add DeskFerry?
Yes. A common split is Make for fixed data movement and DeskFerry for the steps where something has to be read, judged or written.
Do I need to rebuild my scenarios?
No. Nothing is imported. You describe the outcome of a scenario, and the agent does the job its own way. Most teams move a few judgment-heavy scenarios and leave the rest.
Turn your busiest router into one paragraph
Write the routes down as a brief, connect the same apps, and see which invoices the agent holds and why.