Make in depth

How Make scenarios are built and billed: modules and bundles, mapping functions, routers and filters, iterators and aggregators, data stores, connections and credits.

Scenario

Also called Make workflow, scheduling

What: A Make workflow: a chain of modules on a canvas, started by a trigger module, a webhook or a schedule. Each run is kept in the scenario history with the data every module saw.
Why: The schedule sets how often the scenario runs and so how many credits it uses; the history is where you debug.

Watch out

Scheduling every minute for data that changes once a day.

Modules and bundles

Also called bundle, operation

What: Each module outputs bundles, one per record, and the next module runs once for every bundle it receives.
Why: Bundle counts explain both behaviour and cost: a search returning 30 bundles makes the next module run 30 times.
Example: Search Rows returns 12 rows, so the following Send Email module runs 12 times and uses about 12 credits.

Watch out

Not checking how many bundles a search returned before adding an action after it.

Mapping and functions

Also called formulas, formatDate, ifempty, mapping panel

What: Mapped values appear as pills such as 1.email (module 1, field email). Functions such as lower, trim, ifempty, formatDate, parseDate, get and map transform them; arguments are separated by semicolons.
Why: Functions clean data inside a field without extra modules, which keeps scenarios short and saves credits.

Code

{{lower(trim(1.email))}}
{{ifempty(1.firstName; "there")}}
{{formatDate(1.createdAt; "YYYY-MM-DD"; "Asia/Karachi")}}
{{length(1.items)}}

Watch out

Using commas between function arguments instead of semicolons.

Routers and filters

Also called router, filter, fallback route

What: A router splits a scenario into routes; a filter on each route lets through only bundles that meet its conditions. One route can be marked as the fallback for bundles that match none.
Why: Routers let one scenario handle several cases, and filters stop irrelevant bundles early so later modules do not run or cost credits.

Watch out

Overlapping filters that send one bundle down two routes, so a customer gets two emails.

Iterators and aggregators

Also called iterator, array aggregator, text aggregator

What: An iterator splits one bundle holding an array into many bundles; an aggregator collects many bundles back into one, such as an array, a text block or a CSV. Each aggregator has a source module where grouping starts.
Why: They are how Make handles lists such as order line items, and forgetting one is the most common cause of duplicate messages.
Example: Order with 8 items: Iterator over items, then Text aggregator, then one Send Email with the full list.

Watch out

Choosing the wrong source module, so the aggregator groups the wrong bundles.

Data store

Also called Make database, key-value store

What: A small built-in table in Make with a key per record and a data structure you define, which scenarios can add to, search, update and delete from.
Why: Useful for state between runs, such as processed IDs or the last sync time. Size depends on plan, so it is not a replacement for a client CRM or database.

Watch out

Keeping important client records only in a data store the client cannot easily see or back up.

Connections and the HTTP app

Also called connection, HTTP module, custom webhook, Make Code app, Code app

What: Connections store access to apps through OAuth or keys. The HTTP app calls any API, custom webhooks receive data from any app, and on paid plans the Code app runs JavaScript or Python.
Why: These reach apps with no Make module. Whose account a connection uses decides who controls the workflow later.

Watch out

Building a client scenario on your personal Google connection.

Credits

Also called operations, Make pricing

What: The billing unit on Make. As of 2026-09 the pricing page says most module actions use 1 credit; the free plan has 1,000 a month and two active scenarios, and the Code app uses credits per second of run time.
Why: Estimating runs per month times modules executed per run, plus a margin, tells a client what their plan must cover.
Example: A 5-module scenario every 15 minutes processing one record per run: about 2,900 runs, roughly 14,000 credits a month.

Watch out

Forgetting that the trigger check itself uses a credit even when nothing is new.

Go deeper