Make

Make is a visual workflow automation platform that helps teams connect apps, transform data, and orchestrate multi-step scenarios — including AI agents and MCP tools — with granular control. In 2026 it ships next-gen AI Agents built directly into the Scenario Builder, a Make Grid observability map, and MCP server support on every plan.

What sets it apartA visual AI automation canvas where next-gen Make AI Agents, MCP, and Make Grid observability ship together on every paid plan — purpose-built for complex, multi-app workflows.

Quick Verdict

Make (make.com) is the visual AI automation platform I keep coming back to when automations stop being simple. It’s where I go when Zapier feels too rigid and n8n feels too raw. In 2026 it’s also become one of the cleanest places to build and run AI agents without bolting on a separate tool.

If your workflows branch, loop, transform messy data, or need an AI step in the middle, Make is the shortest path from idea to production I know of. If all you need is one trigger firing one action, the canvas might feel like overkill — but the free plan costs you nothing to try.

What Make Is (and Isn’t) in 2026

Make is a visual workflow automation platform — originally launched as Integromat in 2012, renamed Make in 2022, and now part of Celonis. You build “scenarios” on a drag-and-drop canvas where each module is an app action, an HTTP call, an AI step, or an agent decision.

As of July 2026, Make describes itself as a “visual AI automation platform” connecting 3,000+ apps, with three big additions since 2024: Make Grid for live landscape observability, Make AI Agents that live inside the scenario canvas, and the Make MCP Server that turns any scenario into a tool any AI can call (source: make.com, make.com/en/ai-agents).

What Make isn’t. It’s not a code-first orchestration framework. If you want to write Python agents from scratch, LangChain LangGraph or CrewAI are closer fits. It’s also not a pure chatbot builder — though its Library of Agents includes ready-made chatbots for HR onboarding and Slack search.

“Make is a leading integration and automation development platform which empowers businesses across all verticals to visualize systems, streamline processes and put AI to work.” — Make’s own About page, July 2026.

Quick definitions you’ll see throughout:

• Scenario: an automated workflow you build in Make.
• Module: a single step (app action, HTTP call, AI call, router, filter, etc.).
• Operation: a single module run that processes one bundle of data — the unit Make counts for billing.
• Bundle: a unit of data passed between modules — think of one email, one row, one file.
• Credit/Operation: on Make’s pricing page this is the unit that gets consumed; one operation ≈ one credit for most modules.

Who Should Use Make in 2026

Make fits four broad buyer profiles well, and a fifth one not at all.

Pick Make if you are:

1. An ops or RevOps team stitching CRMs, spreadsheets, support tools, and AI together without waiting on engineering. Perks’s 1,800-person global team uses Make this way (source: make.com success story, March 12, 2026).
2. An agency or consultancy running automations for many clients. Make’s team workspaces, scenarios-as-templates, and Make Grid map help you manage dozens of client setups at once.
3. A technical founder or developer who wants visual orchestration plus escape hatches (JavaScript/Python via Make Code, raw HTTP, MCP, sub-scenarios, custom functions).
4. A mid-market to enterprise team that needs AI agents grounded in real systems — CRM records, support tickets, internal docs — without sending data to a black-box vendor. Celonis itself replaced a third-party auditing tool with a Make + Vertex AI build and cut AI processing costs by 99.7% (source: make.com, November 25, 2025).

Skip Make if:

• You want a single self-hosted, code-only framework. n8n is the better match.
• You only need one trigger firing one action. Zapier’s free tier will probably feel faster.
• You’re a Fortune 500 with strict on-prem requirements beyond what the on-prem agent offers — that’s where Workato or Tray.io compete.

Latest Features & 2026 Updates

Make has shipped a lot in the last 18 months. Here’s what actually matters in July 2026.

1. Next-gen Make AI Agents (Feb 11, 2026). Agents are no longer a side tool. They live inside the Scenario Builder alongside deterministic modules. You get a step-by-step reasoning panel that shows every decision the agent made, which tools it called, and why. There’s an in-canvas chat for refining instructions, and multi-modal support means agents can accept PDFs, images, and CSVs as input and produce them as output (source: make.com blog, Feb 11, 2026).

2. Library of Agents. A curated collection of pre-built agents you can deploy and adapt — Market Research, Sales Coach, ICP Content Alignment, Lead Qualification, Invoice Extractor, Resume Analysis, HR Onboarding Q&A, and more. Each combines deterministic steps with agentic reasoning so you can see how production-grade agents are structured (source: make.com/en/ai-agents-library).

3. Make MCP Server and MCP Client. MCP (Model Context Protocol) is the open standard for AI-to-tool communication. Make publishes both ends: the Server exposes your scenarios as tools that ChatGPT, Claude, and Cursor can call, and the Client lets your Make scenarios call external MCP servers. Both are free on every Make plan (source: make.com/en/mcp).

4. Maia by Make. Announced at Waves ’25 (October 16, 2025), Maia is Make’s natural-language builder. You describe an automation in chat, and Maia drafts the scenario while keeping full visual transparency so you can verify every step. Think of it as Make’s answer to vibe-coding, but for automation (source: make.com, Oct 16, 2025).

5. Make Grid. Live visual map of every scenario, agent, and data flow in your Make organization. Launched in closed beta in November 2024 and now GA on every paid plan. ChargeGuru reports saving 1–2 hours per day on debugging because they can spot failing Salesforce integrations in minutes instead of days (source: make.com, June 13, 2025).

6. Module Tools. Turn any Make module into a callable tool that AI agents can use — no need to wrap a whole scenario. Ships on all plans since October 2025.

7. Make Code (JavaScript & Python). Run custom code inside scenarios with sandboxed execution, input mapping, and 30-second runtime (300 seconds on Enterprise with third-party libraries). Available on all paid plans since Waves ’25.

8. If–Else and Merge modules. Native conditional branching. Previously you needed workarounds. Now first-class modules available across all plans since early 2026.

9. Subscenarios and Scenario Sharing. Chain scenarios together and share them via a single link — making team-wide reuse and community contributions actually practical. Subscenarios shipped in December 2024.

10. Scenario Run Replay. Re-run any past scenario execution using its historical trigger data. Great for backfilling or fixing bugs without breaking production. Available since Waves ’25.

Pricing Breakdown July 2026

Make runs on a credit (operation) model. One module that processes one bundle costs one operation. Error handlers and routers don’t consume credits. AI steps typically cost more depending on token usage.

Here’s what Make’s pricing page shows in July 2026 (source: make.com/en/pricing):

Plan	Price	Operations/mo	Active scenarios	Min interval	Max run time	Key extras
Free	$0	1,000	2	15 min	5 min	No time limit, 3,000+ apps, routers & filters, community support
Core (Make Plan)	$9/mo annual ($13 monthly)	5,000	Unlimited	1 min	40 min	Teams, custom variables, 300+ Make API endpoints, full-text log search
Pro	Higher tier	Higher	Unlimited	1 min	40 min	Priority execution, larger file size, advanced features
Teams	Higher tier	Higher	Unlimited	1 min	40 min	Shared development, team roles, audit logs
Enterprise	Custom	Custom	Unlimited	1 min	40 min	SSO, on-prem agent, custom functions, 24/7 support, 99.5% SLA

Free plan perks that people miss:

• No time limit. You can stay on Free forever.
• 3,000+ apps are available, not a stripped catalog.
• All paid AI features like Make AI Agents and Make MCP Server are not locked behind paywall — though execution will consume credits.

Annual billing saves 15% or more versus monthly. Extra credit bundles (1,000 or 10,000) can be added if you run out mid-cycle, and an auto-purchase option kicks in at 90% usage on Core, Pro, and Teams (source: make.com/en/pricing FAQ).

Pull quote: “The free plan isn’t a trial. It’s the on-ramp.” — paraphrasing what Make’s own product page implies when it says “no time limit on the Free plan.”

Make vs Alternatives (July 2026)

This is the comparison I wish someone had handed me when I was choosing. Pricing and feature claims are from each vendor’s public pages in July 2026 unless noted.

Capability	Make	Zapier	n8n	Workato	Tray.io
Visual builder	Drag-and-drop canvas	Linear step list	Node-based, partly visual	Recipe UI, code-friendly	Visual + code
Apps/modules	3,000+	8,000+	400+ (self-hosted unlimited via HTTP)	1,000+ enterprise apps	700+ connectors
AI agents built-in	Yes (all paid plans)	Yes (Zapier Agents, separate pricing)	Yes (LangChain nodes, code-first)	Workato AI, paid add-on	Merlin AI, enterprise tier
MCP support	Server + Client, all plans	Limited	Yes (community)	Yes	Yes
Pricing start	$9/mo (Core, 5k ops)	$19.99/mo (750 tasks)	Free self-host; Cloud from €20/mo	Enterprise quote	Enterprise quote
Free tier	Yes, no time limit	Yes, 2-step Zaps only	Yes, self-host unlimited	No	No
Self-host option	No (on-prem agent only)	No	Yes (fair-code)	No	No
Best fit	Visual-first mid-market ops, AI agents	Mainstream SMB, broadest connector catalog	Developer-led teams, self-host	Large enterprise recipes	Enterprise iPaaS

Two takeaways from this table:

• Make wins on visual AI agents + value. Make AI Agents, Make Grid, and Make MCP all ship on every paid plan. Zapier charges separately for AI agents.
• n8n wins on self-host and code flexibility. If you need the control plane on your own infrastructure, n8n is the realistic choice.

“We’ve seen many Make users switch from Zapier after finding Make to be a much more cost-effective AI automation solution.” — Make, April 30, 2026 comparison post.

That quote is from Make itself, so take the framing with a grain of salt — but the underlying cost math is real. Make’s Core plan at $9/month for 5,000 operations is meaningfully cheaper than Zapier’s $19.99/month for 750 tasks, even before AI features enter the picture.

Real-World Use Cases

These are real customer stories Make published in 2025–2026. I pulled the numbers from each success story page.

1. Stellantis &You UK — auto-closing 18,000 customer messages. A UK auto-dealership group running 700+ Make scenarios built a two-way SMS + AI sentiment workflow. Over 12 months it analyzed 47,000 messages, auto-closed 18,000, and saved 151 hours of manual triage (source: make.com, May 12, 2025).

2. Celonis — 99.7% reduction in expense auditing costs. Celonis’s own finance team replaced a five-figure annual third-party auditing tool with a Make + Vertex AI + monday.com build. They now process ~1,000 documents per dollar and trust the agent enough to start letting it auto-reject obvious violations after just 30–40 days in production (source: make.com, November 25, 2025).

3. FranklinCovey — $100,000s saved, 30 days to 2 hours. The training and consulting firm replaced a 15-year-old service bus and avoided a $60K/year finance app by building custom Make scenarios. One HR process went from 30 days to 2 hours (source: make.com, September 30, 2025).

4. Perk (formerly TravelPerk) — empowering non-technical teams. A 1,800-person global company used Make to let non-engineering departments own their own operational workflows, with “Automation Engineer” ambassadors embedded in each department (source: make.com, March 12, 2026).

5. ChargeGuru — 2 hours saved per day, week to 15 minutes. The EV charging company uses Make Grid + Custom Functions to debug faster and cut lead import setup from a week to 15 minutes (source: make.com, June 13, 2025).

6. Spanish farmer invoicing — 15 minutes to 20 seconds. A community-challenge winner built a Telegram/WhatsApp AI agent that turns voice messages into VAT-compliant PDFs and logs them to Google Drive. Now used by farmers, freelancers, and construction firms via billfast.ai (source: make.com, September 5, 2025).

7. Globant — 30,000+ employees. A 2026 success story describes how Globant’s Head of Innovation scaled Make across a 30,000-employee consultancy (source: make.com, March 9, 2026).

Limitations & Gotchas

Make is excellent at what it does, but it has edges worth knowing about.

1. Credit math is real. Every operation counts. A scenario that loops 5,000 rows with 8 modules per row is 40,000 operations — not 8. The Make Help Center has a full operations explainer; their nudge system warns you at 75% and 90% usage (source: help.make.com/operations).

2. The free plan only gives you 2 active scenarios. That’s fine for learning, limiting for production. Plan to upgrade once you’re past prototype.

3. Steeper learning curve than Zapier. Routers, iterators, aggregators, and bundles take longer to internalize than Zapier’s linear steps. The Make Academy has 38 interactive courses, and the Community forum has 45,000+ registered members to help.

4. No self-hosting. If your security policy requires the control plane on your own infrastructure, n8n is the realistic alternative. Make offers an on-prem agent for Enterprise customers to bridge local systems, but Make itself runs in AWS.

5. AI agent behavior is still non-deterministic. Even with the new reasoning panel, agents can surprise you. Use deterministic modules for the parts that must be predictable, and agents for the parts that need judgment. The Make team’s own guidance is to “let deterministic steps handle the control flow. Let agents handle judgment and selection” (source: make.com blog, February 11, 2026).

6. Custom pricing above 1M operations. Enterprise contracts kick in once you’re past a million monthly operations. If your workload crosses that, expect a sales conversation.

7. AI pricing is layered. Using Make’s built-in AI Provider counts operations differently than bring-your-own-key. Bring-your-own (Anthropic, OpenAI, Gemini) gives you cost control; built-in AI Provider is simpler but less predictable.

Privacy, Security & Compliance

Make has invested heavily here, which matters because AI agents now touch customer data.

• SOC 2 Type II and SOC 3 — both completed and publicly summarized (source: make.com/en/security; trust.celonis.com).
• ISO 27001 certified for the information security program (source: make.com/en/security; make.com/en/enterprise).
• GDPR compliant with a dedicated privacy notice and DPA (source: make.com/en/privacy-notice).
• Data at rest: full-disk AES-256 encryption with AWS KMS-managed keys (source: make.com/en/security).
• Data in transit: TLS 1.2 and 1.3 with AES 256.
• Hosting: AWS, EU and North America regions. Enterprise customers run in a separately managed AWS environment.
• SSO: OAuth2 and SAML2 on Enterprise, plus domain claim to lock down self-service account creation.
• Audit logs and role-based access on Enterprise plans.
• Uptime SLA: 99.5% on Enterprise.
• On-prem agent for securely bridging to local systems like SAP without exposing them to the public internet.
• Pentesting and SAST: Make states it runs regular third-party penetration tests and OWASP-aligned Static Application Security Testing (source: make.com/en/security).

For regulated buyers, the practical checklist is short: confirm your data residency (EU vs US), confirm whether you need SSO + domain claim (Enterprise), and confirm whether the on-prem agent covers your local systems.

FAQ

What is Make.com in 2026?

Make.com is a visual AI automation platform. You design scenarios on a canvas that connect 3,000+ apps, transform data, and run AI agents built directly inside the same canvas. Make is owned by Celonis, the German-American process intelligence company valued at an estimated $11–13 billion as of 2026 (source: Wikipedia/Celonis; make.com).

How much does Make cost in July 2026?

Free plan: $0 for 1,000 operations per month with no time limit. Core (Make Plan): $9/month on annual billing for 5,000 operations. Pro and Teams tiers are higher. Enterprise is custom. Annual billing saves 15% or more versus monthly (source: make.com/en/pricing, July 2026).

Is Make cheaper than Zapier?

On comparable workloads, yes. Make’s Core plan starts at $9/month for 5,000 operations. Zapier’s paid plans start at $19.99/month for 750 tasks. Make also includes AI agents and Make Grid on every paid tier, while Zapier sells AI agent features separately. A comparable fully-equipped setup on Zapier can cost over six times more than Make’s Core plan (source: make.com/en/blog/make-vs-zapier, April 30, 2026).

What are Make AI Agents?

Make AI Agents are scenario-aware AI decision modules that live inside the Scenario Builder canvas. Their next-generation form launched on February 11, 2026, adding a step-by-step reasoning panel, in-canvas chat, multi-modal file support, and shareable agent templates via the Library of Agents (source: make.com/en/blog/announcing-next-generation-make-ai-agents).

Does Make support MCP?

Yes. Make ships both a Make MCP Server (so ChatGPT, Claude, Cursor, and other MCP-compatible AIs can call your Make scenarios as tools) and a Make MCP Client (so your Make scenarios can call external MCP servers). MCP is included on every Make plan at no extra cost — you only pay for the operations the scenarios consume (source: make.com/en/mcp).

Can I self-host Make?

No. Make is a cloud SaaS that runs on AWS in EU and North America regions. Enterprise customers can deploy a Make on-prem agent to bridge local systems into Make scenarios, but the Make control plane is not self-hostable. If you need self-hosting, n8n is the realistic alternative.

What is Make Grid?

Make Grid is a live, auto-generated visual map of every scenario, AI agent, sub-scenario, and data flow in your Make organization. It launched in closed beta in November 2024 and is now generally available on every paid Make plan. ChargeGuru uses it to debug Salesforce breaks in minutes rather than days (source: make.com/en/grid; make.com success story, June 13, 2025).

Is Make good for non-technical users?

Yes for simple automations, with some learning curve for complex ones. The next-gen AI Agents (Feb 2026), Maia natural-language builder (Oct 2025), 7,900+ templates, 38 Make Academy courses, and 45,000+ community members all help non-engineers ramp up. Perk’s 1,800-person global team is a strong case study for non-technical adoption (source: make.com success story, March 12, 2026).

How does Make handle AI costs?

Make AI Agents show execution history with token usage and a cost breakdown per call. You can tune prompts, anchor AI decisions with deterministic modules, or bring your own LLM key to control spend directly. Built-in AI via the Make AI Provider is convenient; BYOK is better when cost sensitivity is high (source: make.com blog, February 11, 2026; make.com/en/mcp).

Is Make an iPaaS?

Make is sometimes called a visual iPaaS (integration platform as a service). It sits between lightweight no-code tools like Zapier and full enterprise iPaaS suites like Workato or Tray.io. For most mid-market teams, that middle ground is the sweet spot.

Final Verdict

Make in July 2026 is the most visually transparent AI automation platform I can recommend without reservations. The combination of visual scenarios, next-gen Make AI Agents with a real reasoning panel, Make Grid observability, MCP server on every plan, and a $9 starting price is genuinely hard to beat.

If your work involves complex, branched, data-heavy automations — or AI agents that need to touch real systems — shortlist Make alongside Zapier and n8n. If you only need two-step Zaps, Zapier’s free tier will feel faster. If you need self-hosting or full code control, look at n8n. For most mid-market teams, Make is the right default in 2026.

“Make is built for what comes next. Powerful platform that handles complex workflows, AI agents that actually work alongside your processes, and a visual environment where nothing is hidden.” — Make, April 30, 2026.

The platform has matured enough that I trust it for production-critical work. And with Celonis’s 2026 IPO reportedly on the horizon (source: Wikipedia/Celonis citing PitchBook), the company behind Make is unlikely to disappear anytime soon.