Best Workflow Automation Software in 2026

In short 8 tools, in order of how well they meet our criteria.

In short We judged each tool on 5 criteria, using only facts from its own public pages.

• Cost at volume: What counts toward the bill: tasks per step, credits per action or one execution per run.
• Plan gates: Which plan you need for multi-step workflows, webhooks and premium apps.
• App coverage: Prebuilt steps for the apps you use, or HTTP requests you build yourself.
• AI and code: AI steps and agents, and JavaScript or Python code steps.
• Control: Self-hosting, run history and limits on users.

In short Best for no-code automation across 9,000+ apps.

• Strengths

  Connects 9,000+ apps; HubSpot, Gmail and Outlook steps work on the Free plan, and only select premium apps need a paid plan. Formatter, Filters, Paths, Delay and Looping steps use no tasks, and triggers and polling never count toward the task limit. AI by Zapier steps cost 1 task on standard models, 3 on advanced and 5 on premium, or 1 task when you connect your own model provider. Zapier MCP is available to all accounts and lets AI assistants run Zapier actions at 2 tasks per tool call. Copilot builds Zaps from a plain-language description, with a daily message limit on Free and no limit on paid plans.

• Limits

  The Free plan allows only two-step Zaps and 100 tasks a month, with no premium apps, webhooks, Filters or Paths. Webhooks and Salesforce are premium apps, so they need Professional, from $19.99 a month billed yearly for 750 tasks. Price climbs with task volume: 10,000 tasks a month costs $129 on Professional or $169 on Team, billed yearly. Zapier Agents is priced apart from Zap tasks: Agents Pro is $50 a month, or $33.33 billed yearly, for 1,500 activities.

• Zapier pricing
• Zapier website
• Zapier pricing page

In short Best for visual scenarios with unlimited users.

• Strengths

  Routers and error handler modules use no credits, and most other module actions use 1 credit. Every plan includes unlimited users, and Teams adds team roles and shared scenario templates from $38 a month billed yearly. AI agents (beta) and the Make MCP server are on every plan, Free included. Maia builds scenarios and agents from a goal you describe in plain language. Core starts at $12 a month billed yearly for 10,000 credits, with unlimited active scenarios and runs as often as every minute. The Make Code app runs JavaScript or Python on paid plans at 2 credits per second of execution, up to 30 seconds a run (300 seconds on Enterprise).

• Limits

  The Free plan allows 2 active scenarios, runs no more often than every 15 minutes and stops a run after 5 minutes. When credits run out, scenarios stop until you buy more or upgrade; overage protection is Enterprise only. Custom JavaScript functions, enterprise apps, the on-prem agent and company SSO are Enterprise only.

• Make pricing
• Make website
• Make pricing page

In short Best for technical teams that self-host.

• Strengths

  Bills per workflow execution, not per step: a run counts once however many steps it has or how much data it processes. Every plan includes unlimited users, unlimited workflows and every integration. The free self-hosted Community Edition includes code steps, HTTP requests, webhook triggers and AI agent nodes. Code steps run JavaScript or Python, and the HTTP Request node calls any REST or GraphQL API, on every plan. Every plan has the AI Agent node, MCP client and server, and human approval before an agent runs a sensitive tool.

• Limits

  Starter allows 2,500 executions a month, 5 concurrent executions and runs of up to 5 minutes. SSO, Git version control and environments start at Business, $800 a month billed yearly, which runs self-hosted only. On n8n Cloud, Python code steps cannot import libraries. The Community Edition license allows only internal business, non-commercial or personal use, and you run and maintain the servers.

• n8n pricing
• n8n review
• n8n website
• n8n pricing page

In short Best for open-source flows at one credit per run.

• Strengths

  A run costs 1 credit however many steps it has; only AI steps cost more. The core is open source under the MIT license and self-hosts with Docker Compose or a Kubernetes Helm chart. The Free plan gives 1,000 credits a month for up to 5 users with no card, including flows, agents, tables and MCP. 768 open-source pieces, including HubSpot, Salesforce, Gmail, Outlook, HTTP and Webhook, come with every plan. Plus and Team are flat prices, $20 and $200 a month, with overage at $0.007 per credit.

• Limits

  AI steps on Activepieces-paid models cost 2 to 20 credits each; using your own AI keys needs Plus, $16 a month billed yearly. Projects, SSO and roles start at Team, $166 a month billed yearly. Git Sync, audit logs, SCIM and custom roles need Ultimate, which is priced by sales. The LinkedIn piece only posts updates to a profile or Company Page.

• Activepieces website
• Activepieces pricing page

In short Best for GTM data enrichment workflows.

• Strengths

  Every plan, Free included, has unlimited seats and tables, multi-provider waterfalls and the Claygent AI research agent. Connects to 150+ data and AI providers, and charges no Actions or Data Credits when an enrichment returns no result. You can bring your own data provider and AI keys: those runs skip Data Credits and use 1 Action each. Launch adds job change, company news and social listening signals, plus an email sequencer that sends from Google Workspace, Microsoft 365 or SMTP accounts. Growth syncs with CRMs and data warehouses, calls any HTTP API and receives data through webhooks.

• Limits

  HTTP API calls, webhooks and CRM integrations start at Growth: $495 a month billed monthly or $446 a month billed annually. Billing runs on two meters: Actions reset each cycle and cannot be topped up, and Data Credit top-ups cost a 30% premium. The Sequencer is email only, with up to 4 emails per sequence and no LinkedIn, dialer or manual task steps. Each webhook source accepts 50,000 submissions; only Enterprise can auto-delete rows so one webhook keeps receiving data.

• Clay pricing
• Clay review
• Clay website
• Clay pricing page

In short Best for enterprise integrations and on-prem systems.

• Strengths

  Lists 10,000+ app and system integrations, with HubSpot and Salesforce connectors that create, update and search records. The Free plan grants 50,000 credits once, with the same platform capabilities as Pro. One credit balance covers recipe tasks, API calls, data pipeline rows and genie actions, and steps skipped by conditional logic are not counted as tasks. The on-prem agent opens a secure connection from a private network to Workato cloud to reach databases, files and apps behind a firewall. JavaScript and Python connectors run custom code inside recipes.

• Limits

  Pro allows up to 3 users per workspace; unlimited users and workspaces need Enterprise, which has no public price. Free credits are a one-time grant, and unused Pro credits expire at the end of each billing month. AI by Workato and on-prem connectivity are available only on specific plans, and AI by Workato needs a signed AI feature addendum.

• Workato website
• Workato pricing page

In short Best for AI agents across business apps.

• Strengths

  Pro costs $37 a month with 20,000 credits, unlimited seats and unlimited teams. 250+ prebuilt connectors, including HubSpot (90 tools), Salesforce (109 tools), Gmail and Outlook. Every agent has a Python code sandbox with full internet access and 80+ preinstalled packages, billed as compute time with no per-call charge. Agents run on schedules, on app events, or from a webhook URL that any service can call with an HTTP POST. Approval settings can pause an agent before actions such as sending email or deleting records.

• Limits

  No free plan: the 14-day Pro trial needs a card and turns into a paid subscription unless you cancel. Agent cost varies by run: model tokens, connector calls, tool charges and compute minutes, plus an 8% orchestration fee on the total. Pro credits do not roll over, and overage is off until you enable it, then capped at 1,000,000 credits ($5,000) per billing period. SSO, SCIM, role-based access control and audit logs are Enterprise only.

• Gumloop pricing
• Gumloop website
• Gumloop pricing page

In short Best for routines written in plain English.

• Strengths

  Routines run on a schedule or on events such as an email received, a Slack message or a meeting ending, and you set them up by describing them in plain English. A dry run fires a routine on real data and shows what it would have done without sending, posting or writing anything. Connects to 1,000+ tools and any hosted MCP server, with HubSpot and Salesforce actions for contacts, deals and tasks. Every seat adds its credits to one workspace pool, and admins can cap how much each teammate uses. No overage bills: when credits run out Lindy pauses, and top-ups cost $10 per 1,000 credits and don't expire.

• Limits

  There is no visual workflow editor: a routine is a trigger, a plain-English prompt and a destination. Every active user needs a paid seat, from $29.99 a month, and no annual price is published. Credit use varies by task, from 2 credits for an everyday ask to 2,500 for a big build, and monthly plan credits don't roll over. LinkedIn actions cover posts, comments and profile lookups, not messages or connection requests, and email goes out through your connected Gmail or Outlook.

• Lindy pricing
• Lindy review
• Lindy website
• Lindy pricing page

In short Ask each vendor on your shortlist these questions before you buy.

• What counts toward the bill: tasks per step, credits per action or one execution per run?
• Which plan do you need for multi-step workflows, webhooks and premium apps?
• Does it have prebuilt steps for the apps you use, or will you build HTTP requests yourself?
• Can it run AI steps and custom code?
• Do you need self-hosting, and how many users does each plan allow?

In short 5 questions buyers ask about workflow automation tools, answered from each vendor's public pages.