Make

Make for AI Workflows: Sequential Automation & AI Modules

Last updated: November 11, 2026 · Reviewed by: IntegrateStack Editorial

Make ↗ gives you two ways to use AI. In a normal scenario, an AI module is one step in a fixed sequence: watch an inbox, classify the email, draft a reply, send it. With the Make AI Agent app, the model decides which tools to call, keeps conversation memory, and can search files you upload. Both run on Make's cloud and connect to its library of more than 3,500 apps.

Make now builds agents as well as workflows

Checked November 11, 2026, against Make's help center and pricing page.

Scenario with AI modules Make AI Agent app
Who decides the steps You, in the scenario design The model, from the tools you give it
Tools The modules in the scenario Modules, whole scenarios, MCP tools and sub-agents
Memory Only what you store yourself (for example, a data store) Conversation memory by conversation ID, with a history limit and compaction
Your documents Passed in as module inputs Knowledge files searched with retrieval (RAG)
Plan All plans All plans with Make's AI Provider; other model providers need a paid plan. Open beta

Sources: Make AI Agent app, Make AI agents, Make pricing.

Use the agent app when steps vary

A scenario is the better fit when every input goes through the same steps, because each run is predictable and easy to debug. The AI Agent app suits work where the next step depends on what the model finds, such as a support agent that sometimes looks up an order, sometimes checks a knowledge file and sometimes hands off to a human. You give it tools (existing modules, other scenarios, MCP tools, or sub-agents), and it remembers earlier turns of the same conversation.

The app is in open beta. If you need to self-host the agent or write custom code inside it, compare n8n's AI Agent node.

Start from a template

Ready-to-import scenarios are in the Make AI agent blueprint library.

Build an AI scenario in Make →

Affiliate disclosure

Make charges per module run

Make bills in credits. Most module actions use one credit, so a scenario that processes 500 items through one AI module uses about 500 credits. Routers and error handlers are free, and some features from Make's AI Provider cost more than one credit per run. The free plan includes 1,000 credits a month, and extra credits are sold in packs of 1,000 or 10,000.

For high-volume AI work, that per-step pricing adds up faster than n8n's, which counts one execution per workflow run regardless of steps.

Frequently asked questions

Can Make build AI agents? ▼

Yes, with the Make AI Agent app (open beta): tools, conversation memory and knowledge files, on all plans with Make's AI Provider.

How does Make pricing work for AI automation at scale? ▼

About one credit per module action; routers and error handlers are free. The free plan includes 1,000 credits a month.

Can I self-host Make? ▼

No. Make ↗ runs only as a cloud service.

Related guides and pages