# Lindy vs n8n: the open-source system you self-host

> Lindy is a hosted AI teammate priced in credits, and n8n is a self-hostable workflow tool billed by executions. This comparison sorts both by what you actually own and run, next to Kortix, the open-source AI Operating System.

URL: https://lindyalternative.com/lindy-vs-n8n

_Compare_

## Lindy vs n8n, sorted by what you own

Lindy rents the runtime; n8n and Kortix both self-host, and every rival cell quotes the project's own pages.

| | Kortix (recommended) | Lindy | n8n |
| --- | --- | --- | --- |
| Open source | Yes. Read it, fork it, audit it. | No. A hosted service. | The code is public and self-hostable; n8n calls its model fair-code. |
| Self-hosting | Kortix Cloud, your VPC, or on-prem. Self-host is free. | Lindy's cloud only. | Self-host the full tool; n8n Cloud is the managed option. |
| Model choice | Any model provider with your own keys. | A model per task from those Lindy hosts. | Connect OpenAI, Anthropic, Google, or open-source models. |
| Where agents and memory live | One git repo you own. | In Lindy's cloud; it says memory sits in plain files. | In your own instance; n8n Cloud keeps data in the EU. |
| Pricing model | Free tier with pooled credits; self-host is free. | Credits per active seat, refreshed monthly, no rollover. | Executions per plan; one run of a whole workflow is one. |
| Extensibility | Agents, skills, connectors, and tools as files in the repo. | 40+ skills and any MCP server. | Custom code in JavaScript or Python, plus npm packages. |
| Approvals | Approval gates you set. Off until you set them. | Anything with outside impact waits for approval. | Human approval steps inside a workflow. |

## Lindy vs n8n, and the one question that sorts them

The pair looks close until you ask who runs it. n8n is a workflow tool you download and run yourself. Lindy is a hosted teammate you reach over the web, and you rent its runtime by the credit. Kortix is a third option: the open-source AI Operating System, where agents, skills, memory, and connectors sit in one git repo you own and the same stack self-hosts.

## What n8n is

n8n's own docs call it a fair-code workflow automation tool that combines AI features with business process automation. Its repository describes "a fair-code platform to build and deploy AI agents and workflows," with a visual canvas, custom code, 1,500+ integrations, and a self-hostable server ([n8n on GitHub](https://github.com/n8n-io/n8n), checked October 2026). It bills n8n Cloud by executions, where one run of an entire workflow is one execution. Its model layer connects OpenAI, Anthropic, Google, or open-source models.

## What Lindy is

Lindy is a hosted AI teammate: it plugs into your tools, learns how you work, and takes work off your plate, in its own words. You reach it over the web, Slack, email, and SMS, with 1,000+ integrations and 40+ skills. It bills in credits that refresh monthly and do not roll over, and every active seat draws from one shared pool ([Lindy pricing](https://www.lindy.ai/pricing), checked October 2026). The credit arithmetic is on the [Lindy priced out](/lindy-pricing) page.

## Where the system lives, and how to run it yourself

This is the line that sorts the three. n8n self-hosts on a server you control, and so does Kortix. Lindy does not: its own pages show no self-hosting option, so its runtime stays in Lindy's cloud. On Kortix the whole system is files. Each agent is `agents/<name>.md` plus an `agents.<name>` block in `kortix.yaml`, and company memory is plain files today in the same repo. Three commands install and start the stack on a Linux box you control:

```sh
curl -fsSL https://kortix.com/install | bash
kortix self-host init --domain kortix.example.com --admin-email you@example.com
kortix self-host start
```

The steps are in the [Kortix docs](https://kortix.com/docs/host), and the [self-hosting guide](/self-hosting) walks through the choices. Both self-hostable options hand you the runtime, and only Kortix hands you the configuration and the memory as files you can read, fork, and audit. The wider field is on [Lindy alternatives](/lindy-alternatives), and if the other tool on your desk is Zapier, [Lindy vs Zapier](/lindy-vs-zapier) covers that pair.

_FAQ_

## Lindy vs n8n, answered

Short answers about self-hosting, models, and where the memory lives.

### Can I self-host both Lindy and n8n?

n8n self-hosts on a server you control, and Lindy does not: its own pages show no self-hosting option. Kortix self-hosts too, and the stack is free to run on your own machines.

### Which models can each one use?

n8n connects OpenAI, Anthropic, Google, or open-source models. Lindy offers a model per task from those it hosts. Kortix runs any model provider with your own keys.

### Where do the agents and the memory live?

On n8n they live in your instance when self-hosted; on Lindy they stay in Lindy's cloud. On Kortix they are files in one git repo you own.

### Is n8n open source?

n8n describes its own model as fair-code, with source you can read and self-host. Kortix is open source: you can read it, fork it, and audit it. For the wider set, see [Lindy alternatives](/lindy-alternatives).

## Run the stack on your own machines.

Self-host is free, and the configuration stays yours.

[Try Kortix](https://kortix.com) · [See the self-hosting guide](https://lindyalternative.com/self-hosting)

Open source · Any model, your keys · Self-host, VPC, or on-prem