How much does Make cost in 2026?

Make (formerly Integromat) changed its billing unit from operations to credits on 27 August 2025. As of June 2026, one credit equals one module action — creating a record, fetching an email, transforming data. Routers and error handlers cost no credits, and AI modules charge a variable number of credits based on the tokens, files, or pages they process.

Make pricing tiers (as of June 2026)

Annual billing reduces these rates by roughly 15%. Credit allotments scale within a tier — you buy additional credits as volume grows — and a November 2025 adjustment raised the cost of extra credits by about 25%, so a scenario that runs hot should be sized against its credit ceiling rather than the headline price.

What actually consumes credits

The credit model rewards efficient scenario design and penalises wasteful module use, which is the opposite of a flat per-task plan.

Diagram source (Mermaid)

flowchart TD
  A[A step runs in a scenario] --> B{What kind of module?}
  B -- App action: create/fetch/update --> C[1 credit]
  B -- Router / error handler --> D[0 credits]
  B -- AI module --> E[Variable credits by tokens/files/pages]
  E --> F[Can cost several times a normal module]

AI modules are where credits disappear fastest: a single AI summarisation or extraction step can consume several times the credits of an ordinary module, so an AI-heavy scenario that looks cheap on the tier price can drain its monthly allotment well before month end.

How Make compares

Make's credit model sits between Zapier's task-based billing and the per-execution billing of n8n. For high-volume, many-module scenarios it can be more economical than Zapier; for simple two-step automations Zapier's free and entry tiers are often simpler to reason about. Teams that need to keep all execution in their own infrastructure usually compare against self-hosted n8n instead.

Editor's Note: Across ShadowGen Make builds in 2024-26, the August 2025 switch to credits was the change clients felt most. On one content-operations scenario an AI summarisation module burned credits several times faster per run than the surrounding modules, so the scenario blew through its monthly credits two weeks early. The practice we apply now: meter AI modules in a short pilot, then size the plan against that measured burn rather than the module count. — Rafal Fila, ShadowGen