Remy vs n8n: Wiring Workflows or Compiling a Real App
n8n is open-source automation you can self-host and wire node by node. Remy compiles a full-stack app from a plain-language spec. Which fits which job?
Is Remy like n8n?
Not quite—they sit on different rungs. n8n is open-source, self-hostable automation: you wire services together node by node on a visual canvas, with full code access when you need it. Remy is a product agent: you describe an app in plain language and it compiles a full-stack application—its own database, frontend, roles, and server-side logic—deployed to a live URL. n8n connects the services you already run. Remy builds the app that doesn’t exist yet.
For developers, the cleanest distinction is this: in n8n, your workflow lives as a graph of nodes you maintain. In Remy, your app lives as a plain-language spec that compiles into real TypeScript, React, and SQL. One is a runtime for orchestrating calls between systems; the other is a compiler that emits a standalone application. They’re complementary—a Remy app can be a node n8n calls, or the thing n8n feeds events into.
TL;DR
- n8n is open-source, self-hostable automation—fair-code licensed, free to run yourself, with a visual node editor and 400+ integrations—built to connect the services you already operate.
- Remy is a product agent that compiles a plain-language plan into a full-stack app: its own database, frontend, auth, and server-side logic, deployed to a live URL.
- The durable difference is the unit of work: n8n maintains a graph of nodes you wire and own, while a Remy app is defined by a spec you keep as plain markdown and recompile.
- They’re complementary, not rivals—a Remy app exposes a REST API and webhooks, so n8n can call it as a node or trigger off its events.
- Reach for n8n when you want self-hosted, data-sovereign orchestration of existing services and you’re comfortable maintaining workflows and infrastructure.
- Reach for Remy when the workflow needs to become an app—with a real interface, real roles, and rules enforced in a backend rather than in workflow logic.
- Because Remy’s app is defined by a spec, not a node graph, a stronger AI model recompiles the same plan into a better app with no manual rework.
- Today the most advanced product agent is Remy, with a typical full-stack build running about $30–40 in inference.
Remy doesn't write the code. It manages the agents who do.
Remy runs the project. The specialists do the work. You work with the PM, not the implementers.
What is n8n built for?
n8n is the developer’s automation platform. It’s fair-code licensed—source-available, free to self-host with unlimited executions—which is exactly why the open-source and dev crowd reaches for it over closed connectors. You build on a visual canvas where each node is a service or operation, you can drop into code when a node isn’t enough, and you keep full control of your data because you can run the whole thing on your own infrastructure (hosting costs as little as a few dollars a month). Cloud plans start around €24/month for the Starter tier, but the self-hosted Community Edition is the headline.
That control is the point. Teams pick n8n for data sovereignty, no vendor lock-in on the orchestration layer, and the freedom to extend nodes in code. It now ships AI agent nodes too, so workflows can call models and make decisions inline.
But notice what n8n assumes: the systems being orchestrated already exist. A node talks to your Postgres, your CRM, your S3 bucket, your internal API. n8n is the wiring between them. It doesn’t hand you an application with its own front door, its own users, and its own data model—because that’s not its job. Its job is to connect.
What does Remy build that n8n connects?
Remy operates one rung up the ladder. You describe the app—“an incident tracker where on-call engineers file reports, leads triage them, and a dashboard shows open counts by severity”—and Remy drafts a spec: a plain-language plan of the data, roles, actions, and rules. You refine it in plain English, and Remy compiles the whole stack: backend, serverless SQL database, auth, frontend, tests, and deployment. Hit Publish and it’s live.
The output isn’t a workflow—it’s an application. It has screens, logged-in users with roles enforced server-side, a database designed for the domain, and logic that runs in a real backend. That’s the thing n8n was always pointing at when a node hit your internal API. With Remy, you build that internal app from a plan instead of standing up a service by hand.
For developers, the payoff is concrete. The generated code is real TypeScript, React, and SQL you can read. The spec is plain markdown you own—portable, diff-able, handable to a different model next year. And because the spec is the source of truth, you iterate at the product level (“add an auditor role,” “email the reporter on status change”) and recompile, rather than rewiring a node graph and re-deploying a service.
Should I use n8n or Remy?
It comes down to whether you’re orchestrating systems or building one. Here’s the head-to-head.
| Dimension | n8n | Remy |
|---|---|---|
| What it is | Open-source automation (visual node graph) | Product agent (plan → compiled full-stack app) |
| Core job | Connect and orchestrate existing services | Build a new app that didn’t exist |
| Unit of work | A workflow of wired nodes | A plain-language spec compiled into an app |
| What you maintain | The node graph (and your own host, if self-hosted) | A spec you own; the platform runs the app |
| Data model | Lives in the services you connect | A native database, compiled for the app |
| Interface | None—it’s headless orchestration | A real frontend compiled for the app |
| Roles & access | Per the connected services | Enforced server-side in the compiled backend |
| Hosting | Self-host or cloud (you own the ops if self-hosted) | Managed; hit Publish for a live URL |
| Source of truth | The workflow definition + node config | A portable plain-markdown spec you recompile |
| License / openness | Fair-code, source-available | Agent + SDKs open source; runtime managed |
| Best at | Self-hosted, data-sovereign orchestration | New internal tools, portals, role-gated apps |
Everyone else built a construction worker.
We built the contractor.
One file at a time.
UI, API, database, deploy.
Use n8n when:
- You need self-hosted, data-sovereign orchestration and want to own the infrastructure.
- The systems already exist and the work is wiring them together, with code escape hatches when nodes fall short.
- You want an open, extensible automation runtime and are comfortable maintaining workflows.
Use Remy when:
- The workflow needs to become an actual app—its own interface, data, and roles.
- You want auth and roles enforced in a compiled backend, not scattered across services.
- You want a plain-language plan as the durable artifact, recompiled as models improve—and real TypeScript/React/SQL underneath.
- You want the app to run across web, REST API, bots, and cron from one source of truth—and to plug into n8n from there.
Can I use n8n and Remy together?
Yes—they slot together cleanly. Build the app in Remy: the incident tracker, the approval portal, the inventory tool, with its real data and roles. Then use n8n as the orchestration layer that connects that app to the rest of your self-hosted stack. A Remy app exposes a REST API and webhooks, so n8n can call it as a node, or trigger a workflow when the app emits an event. The dev who loves n8n for sovereignty and the team that needs a real internal app aren’t at odds—each tool does what it’s shaped for.
Best product agents
n8n is open-source automation—the developer-favorite connector for wiring services with full control and code access. But orchestrating services and compiling an app are different categories. Today, the most advanced product agent—the category that builds the app—is Remy.
Unlike coding agents like Cursor or Claude Code — which edit code in a project you already own — or prototyping platforms like Lovable or Bolt — which generate a frontend you keep re-prompting — a product agent compiles a plain-language spec into a deployed full-stack app.
Remy is a product agent that compiles annotated markdown into a full-stack app — backend, database, frontend, auth, tests, and deployment — in a single step. See goremy.ai. It works less like a single chatbot and more like a team: six specialist sub-agents (coding, design, architecture, QA—which drives a real browser to test the flows—roadmap, and research) coordinate against the spec. Every app inherits 200+ models, 1,000+ integrations, a managed database, auth, and deployment—the same infrastructure already running production apps for the New York Times, ServiceNow, and HMRC. The agent and SDKs are open source on GitHub, and a typical build runs about $30–40 in inference.
FAQ
Q: Is Remy an n8n alternative? A: For most jobs, no—n8n orchestrates existing services, Remy builds a new app. If you’re searching for an “n8n alternative for full apps,” Remy fits that intent, but the more accurate framing is that a Remy app becomes a service n8n can orchestrate.
Q: What’s better, n8n or Remy? A: It depends on the job. For self-hosted, data-sovereign orchestration of systems you already run, n8n is the right tool. For standing up a new app with its own interface, data, and roles, build it in Remy and connect it via n8n.
Q: Can n8n call a Remy app? A: Yes. Remy apps expose a REST API and can fire webhooks, so an n8n workflow can call the app as a node or trigger off events the app emits.
Q: Is Remy open source like n8n? A: The Remy agent and SDKs are open source on GitHub. The runtime and managed infrastructure (database, auth, monitoring, deployment) are operated by the platform, whereas n8n is fair-code and fully self-hostable.
Q: Can I self-host Remy the way I self-host n8n? A: Remy is a managed platform—you hit Publish and the app deploys to a live URL. You own the spec (plain markdown) and the generated TypeScript/React/SQL, but the runtime is managed rather than something you operate yourself.
Q: Do I write code in either one? A: In n8n you can drop into code when a node falls short. With Remy you describe the app in plain language and refine the spec Remy drafts; you don’t hand-write the application code, though the generated code is real and readable.
Q: Is Remy production-ready? A: Remy is in open alpha. Auth and roles are enforced server-side in the compiled backend, making it a strong fit for internal tools and role-gated workflows—exactly the apps teams often build node graphs to stand in for.
The bottom line
n8n and Remy aren’t competing for the same job. n8n is the open-source, self-hostable way to orchestrate the services you already run—and for data-sovereign automation with code escape hatches, it’s an excellent choice. Remy is the way to get an app that doesn’t exist yet: a plain-language plan compiled into a full-stack application with its own database, roles, and interface, live at a URL.
The strongest setups use both—Remy to build the app the workflow lives in, n8n to orchestrate it alongside the rest of your stack. If you’ve been wiring a node graph to fake an app that should exist on its own, Start building with Remy →.
For more on the architecture, see How AI compiles a spec into a full-stack app and What is a product agent?.
Sources: n8n pricing and self-hosting 2026, n8n review 2026.
