Pipedream alternatives after the Workflows shutdown
Pipedream Workflows stops running on March 31, 2027, so every workflow on it needs a new home. The main alternatives are n8n, Zapier, Make, Activepieces, Trigger.dev, Inngest, Kestra and Windmill. They split into hosted no-code tools (Zapier, Make), visual builders you can also host yourself (n8n, Activepieces), code-first engines (Trigger.dev, Inngest, Windmill) and declarative orchestrators (Kestra). Which one fits depends mostly on whether your workflows are built from pre-built actions or from code steps, and whether you need to run them on your own infrastructure.
This page describes each tool from its own documentation and pricing pages, with links, as checked on November 11, 2026. Windmill publishes this page and is one of the options, so the facts about other tools are quoted from their sites rather than summarized from memory. If something here is out of date, write to [email protected].
What Pipedream is shutting down
Pipedream's notice, at pipedream.com/docs/workflows, says:
We want to let you know that Pipedream Workflows and String will shut down on March 31, 2027.
The same notice gives the rest of the timeline:
- Workflows is in maintenance mode: "Your existing workflows will keep running normally through March 31, 2027, but we're no longer accepting signups or subscriptions, and we won't be shipping new features for Workflows."
- "All Workflows data will be permanently deleted by April 30, 2027, including the credentials and OAuth tokens for your connected accounts."
- Subscriptions are cancelled on March 31, 2027 with a prorated refund.
- Pipedream Connect continues: "If you're using our Connect API, MCP Servers, or API Proxy, nothing changes for you."
Pipedream provides an export tool that downloads "the structure of your workflows as JSON, one project at a time". The export does not contain connected-account credentials, so whichever tool you move to, plan to reconnect each account.
If you only use Pipedream Connect to embed integrations in your own product, nothing on this page applies to you.
The alternatives at a glance
| Tool | How you build | Code in a step | Self-hosting | Free option (vendor's wording) |
|---|---|---|---|---|
| n8n | Visual canvas with code nodes | JavaScript, Python | Yes, under the Sustainable Use License | Self-hosted Community Edition; n8n Cloud has a trial |
| Zapier | No-code, with Code by Zapier steps | JavaScript, Python | No self-hosted plan found | "Free forever", 100 tasks per month |
| Make | Visual scenarios | JavaScript, Python (paid plans) | No self-hosted plan found | Free subscription, 1,000 credits per month |
| Activepieces | Visual flows with code steps | TypeScript | Yes, MIT outside the ee folders | "Free forever", 1,000 credits a month |
| Trigger.dev | TypeScript tasks in your codebase | TypeScript, Python via an extension | Yes, Apache 2.0 core | $5 a month of free credits |
| Inngest | Functions in your codebase with an SDK | TypeScript, Python, Go, Kotlin/Java | Yes, server under the SSPL | 50k executions a month |
| Kestra | YAML flows, with a no-code editor | Python, Node.js, Shell, R, Julia, Ruby, PowerShell | Yes, Apache 2.0 | Open-source edition; no public cloud prices |
| Windmill | Scripts, plus a flow editor or workflows as code | TypeScript, Python, Go, Bash and 20+ languages | Yes, AGPLv3 core | Self-hosted Community Edition; Cloud free plan with 1,000 executions per month |
Each alternative
n8n
n8n describes itself as "a workflow automation platform that uniquely combines AI capabilities with business process automation" (n8n.io/llms.txt). Workflows are built on a visual canvas, and the Code node runs "custom JavaScript or Python". Workflows start from a Webhook node or a Schedule Trigger with a custom cron interval, among others. The homepage lists "500+ native integrations" (n8n.io).
On packages, the Code node docs say: "If you use n8n Cloud, you can't import external npm modules", and "On n8n cloud, the Python option for the Code node doesn't allow users to import any Python libraries".
"If you self-host n8n, you can import and use built-in and external npm modules in the Code node", and Python modules have to be included in the n8nio/runners image and explicitly allowlisted.
The self-hosted Community Edition is free; the code is under the Sustainable Use License, with .ee. files requiring an n8n Enterprise license.
n8n Cloud has a trial rather than a free plan, and its Starter plan is "20€ /mo, billed annually" for 2.5K workflow executions (n8n.io/pricing).
n8n is a good fit when your Pipedream workflows are mostly pre-built actions, the people maintaining them prefer a canvas, and you want the option to self-host. Check how your code steps use npm or PyPI packages before moving to n8n Cloud.
Zapier
Zapier calls itself "a no-code AI orchestration platform where teams connect thousands of apps and AI tools" (zapier.com/llms.txt), and its pricing page says it connects "9,000+ apps" (zapier.com/pricing). Code steps are available through Code by Zapier, with "Run Python" and "Run Javascript" actions. Webhooks arrive through a Catch Hook trigger, and the Schedule trigger runs every hour, day, week or month.
The free plan includes 100 tasks per month, and Professional starts "from $19.99 /month" with yearly billing (zapier.com/pricing). We found no self-hosted plan on Zapier's pricing page.
Zapier is a good fit when the Pipedream workflows are built from pre-built actions, the people who maintain them are not developers, and a hosted service is acceptable.
Make
Make describes itself as "a leading visual automation platform for building AI agents, agentic workflows, and integrations without writing code", "connecting 3,500+ pre-built apps" (make.com/llms.txt). The Make Code app adds JavaScript or Python to a scenario and is available on paid plans; importing custom libraries is listed for Enterprise only. Scenarios start from webhooks or on a schedule.
Make's help center gives the standard free subscription as "1,000 credits per month", on its page about cancelling a subscription. Make's pricing page could not be fetched when this page was written, so paid prices are not quoted here. We found no self-hosted plan; Make offers an on-premise agent on Enterprise so cloud scenarios can reach systems on your network.
Make is a good fit for the same profile as Zapier: hosted, visual, built from pre-built apps.
Activepieces
Activepieces describes itself as "AI automation for teams, with governance and security built in" (activepieces.com/docs/llms.txt).
Flows are visual, start from a "Schedule Trigger, Webhook Trigger, or Event Trigger" (docs), and code steps are TypeScript.
In the sandbox mode the docs recommend, code steps run with "no require, no filesystem, no npm"; a process sandbox mode is available "if you genuinely need arbitrary npm packages" (sandboxing docs).
The homepage lists "770+ apps" (activepieces.com).
It can be self-hosted with Docker or Kubernetes, and the code outside packages/ee/ and packages/server/api/src/app/ee is under the MIT Expat license (LICENSE).
The cloud has a free plan with "1,000 credits a month" and a Plus plan at "$20 flat per month" (activepieces.com/pricing).
Activepieces is a good fit when you want a visual builder close to Pipedream's, under a permissive license, on your own servers.
Trigger.dev
Trigger.dev is "the open source platform for building and running durable AI agents and workflows in TypeScript" (docs).
You write tasks "in your existing codebase" and deploy them; Python scripts can run through the pythonExtension build extension.
Tasks can run on a cron schedule, and the docs index lists webhooks as a public beta (llms.txt).
It can be self-hosted, and its docs say its "core is Apache-2.0 licensed". The cloud's free plan includes "$5 / month free credits", and Hobby is "$10 /month" (trigger.dev/pricing).
Trigger.dev is a good fit when your Pipedream workflows were mostly Node.js code steps and you would rather keep that code in your application's repository than in a separate tool. It has no visual builder or pre-built action catalog in the Pipedream sense.
Inngest
Inngest is "the durable workflow engine for AI applications" (inngest.com/llms.txt), with SDKs for TypeScript, Python, Go and Kotlin/Java (README). Functions start from events, a cron trigger or a webhook URL.
The server can be self-hosted, under the Server Side Public License with delayed publication under Apache 2.0; the SDKs are Apache 2.0 (README). The cloud's free plan includes "50k executions", and Pro starts "at $99 /mo" (inngest.com/pricing).
Inngest fits the same profile as Trigger.dev: developers who want durable steps inside their own application code.
Kestra
Kestra is "an open-source workflow orchestration platform. Workflows are defined declaratively in YAML" (kestra.io/llms.txt), and its README says the YAML editor, the no-code editor and the AI Copilot "all edit the same flow".
Script tasks have "dedicated plugins for Python, R, Julia, Ruby, Node.js, Powershell and Shell" (docs), and flows start from schedule or webhook triggers.
Its pricing page lists "2,600+ Plugins".
Kestra is Apache 2.0 and self-hostable, and Kestra Cloud is behind "Request Access" (kestra.io/pricing). Its open-source Pause task waits for someone to click Resume; the pricing page lists Human-in-the-Loop Approvals with custom inputs, and Apps for forms, as Enterprise. Kestra publishes its own Pipedream migration guide.
Kestra is a good fit when the workflows are closer to data or infrastructure pipelines than to app-to-app automations, and your team is comfortable describing them in YAML.
Windmill
Windmill runs scripts, flows, apps and AI agents on workers you host or on Windmill Cloud. A Pipedream code step becomes a Windmill script in the same language: TypeScript, Python, Go or Bash, among more than 20 supported languages. Packages are installed from the script's imports, so a step can use the vendor's SDK from npm or PyPI. Flows start from webhooks, schedules, HTTP routes, emails and other triggers, and can pause on approval steps. Every script also gets an auto-generated form, and apps can be built on top of scripts and flows.
What Windmill has less of: pre-built integrations. Pipedream's workflow docs mention "3,000 integrated apps"; Windmill has resource types and community scripts on its Hub, and for anything else a step calls the API or imports the SDK.
Windmill's core is open source under AGPLv3 and self-hosts with Docker Compose or Kubernetes. Some features are Enterprise Edition only on a self-hosted instance, including schedule error handlers and retries, forms on approval pages, unredacted audit logs, SSO for more than 10 users and Git sync for workspaces with more than 2 users. Windmill Cloud has a free plan with 1,000 executions per month, and the Team plan is $10 per seat per month (pricing).
Windmill is a good fit when your Pipedream workflows are mostly code steps, written by developers, and you want to keep them as plain scripts under version control, with the option to self-host. See Windmill vs Pipedream for the side-by-side and the migration guide below.
How to choose
Answer these in order; the first answer usually narrows the list to two or three tools.
- Do you use only Pipedream Connect? Then nothing changes and you do not need to migrate.
- Must the workflows run on your own infrastructure?
That rules out Zapier and Make, and leaves n8n, Activepieces, Trigger.dev, Inngest, Kestra and Windmill.
Read each license: Sustainable Use License (n8n), MIT with
eefolders (Activepieces), Apache 2.0 (Kestra, Trigger.dev's core), SSPL (Inngest server), AGPLv3 with proprietary Enterprise features (Windmill). - Are your workflows mostly pre-built actions or mostly code steps? Count the code steps in your export. Mostly actions points to Zapier, Make, n8n or Activepieces, which have the largest catalogs. Mostly code points to Trigger.dev, Inngest, Kestra or Windmill, where the step code carries over with the least rewriting.
- Do your code steps import packages? Check each tool's package rules above: several hosted plans restrict or block imports in code steps.
- Who maintains the workflows after the move? Non-developers do better on a canvas (Zapier, Make, n8n, Activepieces). Developers who already keep code in Git do better with Trigger.dev, Inngest, Kestra or Windmill.
- Do you need more than workflows? If people also run the same logic by hand, need approval steps, or need internal UIs, compare what each tool includes for that. On Kestra, the open-source Pause task waits for a Resume click, while approvals with custom inputs and Apps are Enterprise; n8n documents human approval for AI agent tools.
Whatever you pick, run the Pipedream export early: it is one project at a time, and the data is deleted on April 30, 2027.
Moving Pipedream workflows to Windmill
Windmill's migration guide covers the full procedure, including a concept-by-concept mapping and what the export leaves out. In short:
- Export each Pipedream project from its settings.
- Give the guide's prompt to Claude Code, Codex or Cursor, which converts the export into Windmill flows on your machine, with no Windmill account needed.
- Connect a workspace, preview each flow, reconnect the accounts, and deploy with routes and schedules disabled until you switch over.