Pipedream vs n8n(2026)

Pipedream vs n8n: which one should you actually pick?

Pipedream and n8n are the two best picks for developer-leaning teams who've outgrown Zapier — but they solve the problem in opposite ways. n8n is a fair-code, self-hostable visual automation platform that you can run anywhere Docker runs, with a Code node when you need real JS or Python. Pipedream is fully-managed serverless — every step is code (Node.js, Python, Go, or Bash), and you never touch the infrastructure.

The split usually comes down to one question: do you want to own the runtime or not? If self-hosting is a hard requirement (data sovereignty, cost at scale, regulated industry), n8n is the only answer in this pair. If you'd rather never think about Docker containers or scaling, Pipedream wins.

The short version: Pick n8n if you need self-hosting, unlimited executions, or source code you can audit and fork. Pick Pipedream if you want serverless pricing, Pipedream Connect for embedded integrations, or a pure-code workflow where every step is a real function.

When n8n is the right pick

• You need self-hosting. n8n runs on your own VPS, Kubernetes cluster, or anywhere Docker works — no vendor in the middle of your data
• You want unlimited executions at flat cost. Self-hosted n8n on a $10-20/mo VPS can run millions of workflows for the same price
• Open source matters. n8n's code is on GitHub under the Sustainable Use License — not OSI-approved OSS, but you can read it, audit it, and fork it
• You need a visual canvas. n8n's drag-and-drop flow is great for teams who want to see the shape of what's happening
• Regulated industries. Self-hosting + data never leaving your infrastructure is the cleanest compliance story

When Pipedream is the right pick

• You don't want to manage infrastructure. Pipedream handles scaling, retries, and scheduling — you just write code
• Your team writes code anyway. Every step runs full Node.js, Python, Go, or Bash with npm/pip package access
• You're embedding automation in your own product. Pipedream Connect gives you drop-in OAuth, token management, and a component library for customer-facing integrations — n8n has no direct equivalent
• You want serverless pricing. Pipedream charges per credit (1 credit = one second of compute), so sub-second workflows cost basically nothing
• Fast iteration. No Docker builds, no server restarts — push a change and it runs

Pricing — completely different models

n8n's pricing splits in two. Self-hosted is effectively free — just your VPS costs ($10-20/mo on Hetzner or DigitalOcean gets you a lot of runway). n8n Cloud starts around $24/mo (Starter, annual) for 2.5K executions, scaling up from there.

Pipedream has a free tier of 100 daily invocations, which covers most light use. Basic is $19/mo for 10K credits. Credits measure compute time, not runs, so a fast workflow costs way less than the per-step model on Zapier or Make.

The math tilts hard based on volume. A team running 100K executions a month will pay almost nothing on self-hosted n8n but rack up real bills on Pipedream or n8n Cloud. A team running 1K short workflows a month might pay $0 on Pipedream's free tier. The break-even point is usually somewhere around 10-30K executions — below that, Pipedream's managed model is worth the money; above it, self-hosted n8n starts pulling away.

Logic and workflow complexity

Both platforms handle real branching, looping, parallel execution, and error handling. Neither is the "linear steps" trap you see in Zapier.

n8n's approach is visual-first with code when you need it. The canvas handles branching via IF nodes and switches, loops via Split In Batches, and error handling via Error Trigger workflows. When the canvas isn't enough, you drop in a Code node and write JS or Python for that step only.

Pipedream's approach is code-first with visible steps. Every step is a code block, so branching is JavaScript if/else, loops are for loops, and error handling is try/catch. More powerful for devs, less approachable for anyone who can't read the code.

If non-developers need to read the workflow, n8n wins. If developers are the only audience, Pipedream's code-at-every-step model is faster to build and easier to version.

Code and customization

This is the closest fight in the comparison — both give you real code, just at different granularity.

Pipedream gives you a code block at every step. Want to call an internal API, reshape JSON, or pull data from a database inside your workflow? Just write the function. Every npm and pip package is available.

n8n gives you a Code node that you can drop anywhere. It supports JS and Python with a solid stdlib, but not every npm package (n8n runs code in its own VM). For heavy custom logic, you'd reach for HTTP calls to your own service more often than on Pipedream.

Pipedream also ships Pipedream Connect — a full SDK for embedding OAuth flows, token refresh, and a component library inside your own app. If you're a SaaS building customer-facing integrations, Connect is a killer feature. n8n has an embed API but it's nowhere near as mature.

AI and MCP

Both platforms have serious AI stories — this is where they differ less than most comparisons.

• Pipedream: Deep MCP integration as both server and client, native LangChain support inside workflows, and first-class AI agent patterns. You can spin up a tool-calling agent in a single Pipedream step.
• n8n: AI Agent node with LangChain integration, MCP client support, tight bindings for OpenAI, Anthropic, and Ollama. Self-hosted n8n + Ollama is one of the cheapest ways to run local-model AI agents.

If you want the lowest possible cost for AI agents, self-hosted n8n + open-weights models wins on dollars. If you want the fastest developer ergonomics for building agents, Pipedream's pure-code model is tighter.

Compliance and self-hosting

• Self-hosting: n8n yes (VPS, k8s, Docker anywhere). Pipedream no — managed only
• EU data residency: n8n — wherever you deploy the container. Pipedream — US-hosted, sales conversation for residency
• SOC 2: Both are SOC 2 Type II on their hosted offerings
• HIPAA: Neither signs a BAA out of the box; n8n Enterprise self-hosted is the easier path for HIPAA scope
• Air-gapped / on-prem: Only n8n supports this

For regulated industries, there's no contest — n8n's self-host story is the clean answer. Pipedream's managed-only model makes compliance conversations harder.

Migration between them

Pipedream → n8n is harder than most migrations because you're going from pure code to visual-plus-code. A Pipedream workflow with 10 code steps becomes 10 nodes on the n8n canvas — the logic ports cleanly, but you'll want to refactor to use n8n's native nodes where possible instead of dropping everything into Code nodes.

n8n → Pipedream is usually easier. Visual nodes map to Pipedream components, and Code nodes port straight across as Pipedream code steps. The trickier part is re-doing credentials and replacing n8n's queue mode with Pipedream's auto-scaling.

Teams usually move n8n → Pipedream when self-hosting is costing more time than it's saving, or they need Pipedream Connect for embedded integrations. They move Pipedream → n8n when cost at scale becomes a problem, or compliance requires self-hosting.

The honest call

n8n is the right answer if you need self-hosting, want unlimited executions at a flat VPS cost, need to audit or fork the source, have a regulated-industry compliance story, or have a non-developer on the team who'll need to read the canvas.

Pipedream is the right answer if you'd rather never manage infrastructure, your team writes code anyway, you're embedding automation in your own product via Connect, you want serverless per-second pricing, or you're building AI agents as a first-class workload.

Both are production-ready for developer teams, both handle complex workflows, and both skip the "Zapier tax." If visual-first automation for non-developers is what you actually need, look at Make instead. For other matchups, browse Integration Atlas or check the app integration catalog and automation guides.