# Kortix, the open-source alternative to Langflow

> Langflow is an open-source visual builder for AI flows. The alternatives split into visual canvases and code frameworks, and they differ in where your configuration lives and where work runs. Our pick is Kortix, which is open source: your agents, skills, memory and connectors in one git repo you own.

URL: https://langflowalternative.com/

An independent comparison of Langflow alternatives

**The open-source AI Operating System**

`kortix.yaml`

```yaml
kortix_version: 2
default_agent: kortix
agents:
  kortix:
    file: agents/kortix.md
    connectors: all
```

Source: https://kortix.com/docs/project/manifest

Session work reaches `main` through a change request. Merge is default-deny for agents.

[Try Kortix](https://kortix.com) · [Compare the field](https://langflowalternative.com/#compare)

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

_Compare_

## Kortix and the Langflow alternatives, side by side.

Rival columns state what each project documents about itself. The Kortix column quotes Kortix.

| | Kortix (recommended) | Langflow | LangGraph | Dify | n8n | CrewAI | Microsoft AutoGen | FlowiseAI | Dynamiq |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Open source | Yes. Read it, fork it, audit it. | Yes. Python source you can edit. | Yes. Python and TypeScript libraries. | Yes. The community edition self-hosts. | Partly. Public source, with limits on hosting it for others. | Yes. An open-source Python framework. | Yes. Python source, community-managed. | Yes. TypeScript and JavaScript source. | Yes. A Python package. |
| What it is | An AI Operating System: agents, skills, memory, connectors in one git repo. | A visual canvas that builds and deploys agent flows. | A code framework for long-running, stateful agents. | A visual canvas for workflows, RAG and agents. | A visual canvas for fixed-step automation with AI steps. | A Python framework for role-based agent teams. | A Python framework for multi-agent applications, now in maintenance. | A visual builder for agent flows; its repository is archived. | A Python framework for RAG and agent orchestration. |
| Where it runs | Kortix Cloud, your VPC, or your own on-prem network. | Your machine, Docker, or any major cloud. | Your own infrastructure, or the LangSmith platform. | Cloud, your VPC, or self-hosted. | Self-hosted, or n8n Cloud. | Your own Python runtime; AMP adds cloud or on-prem. | Your own Python or .NET runtime. | Self-hosted, or Flowise Cloud. | Your own Python runtime. |
| Models | Any model provider with your own keys. | All major LLMs, vector databases and AI tools. | Any model a LangChain integration supports. | Hundreds of LLMs, proprietary and open-source. | OpenAI, Anthropic, Google, or open-source models. | Models set per agent, in code. | OpenAI and other model clients. | All major LLMs and vector databases. | LLM providers configured in code. |
| Your configuration | Files in a git repo you own. | A project inside the Langflow app. | Python or TypeScript in your repository. | An app inside the Dify dashboard. | Workflows in the n8n editor, exportable as JSON. | Python and JSON files in your project. | Python code in your repository. | Flows inside the Flowise editor. | Python code in your project. |
| How work lands | Session work reaches `main` through a change request. Merge is default-deny for agents. | A flow result returned to the app you build. | Agent state your own code reads. | A response inside the app you built. | A finished workflow run. | Task output your code collects. | A conversation result in code. | A flow result returned to the app you build. | Node output your code reads. |

Rival rows reflect what each project documents publicly as of October 2026, from its own repository or website. Flowise has been archived by its maintainers, and AutoGen is in maintenance mode and points new projects to a successor.

_How Kortix works_

## How Kortix works

Six layers, each one a file or a command you can read.

### 1. One git repo that is the company

Your agents, their skills, your company memory and every connector live in one git repo you own.

`kortix.yaml`

```yaml
kortix_version: 2
default_agent: kortix
agents:
  kortix:
    file: agents/kortix.md
    connectors: all
```

Source: https://kortix.com/docs/project/manifest

### 2. Agents and skills are markdown

An agent is `agents/<name>.md` (behavior) plus an `agents.<name>` block in `kortix.yaml` (governance).

`kortix.yaml`

```yaml
agents:
  support-triage:
    file: agents/support-triage.md   # what it does, in plain markdown
    connectors: all                  # what it may reach
```

Source: README.md in https://github.com/kortix-ai/suna

### 3. Every tool, one governed path

3,000+ apps in a click, plus MCP, OpenAPI, Postman, GraphQL and raw HTTP. Connector credentials are brokered server-side and never enter the machine.

`kortix.yaml`

```yaml
connectors:
  - slug: gmail
    name: Gmail
    provider: composio
    app: gmail
```

Source: https://kortix.com/docs/connect/connectors

### 4. Any model, your keys

Any model provider with your own keys. Or the ChatGPT plan you already pay for. Or sign in with your OpenCode Console account for OpenCode Zen and Go.

`terminal`

```sh
$ kortix providers login openai        # the ChatGPT plan you pay for
$ kortix providers login opencode      # OpenCode Zen
$ kortix providers login opencode-go   # OpenCode Go
```

Source: https://kortix.com/docs/project/models

### 5. A computer per session

One isolated sandbox per session. Each session has its own isolated machine and branch.

`terminal`

```sh
$ kortix sessions new --prompt "Summarize this week's commits and open a change request"
$ kortix cr ls   # review what the agents propose
```

Source: README.md in https://github.com/kortix-ai/suna

### 6. Work lands through a change request

Session work reaches `main` through a change request. Merge is default-deny for agents.

`terminal`

```sh
$ kortix projects clone <project_id> my-project
$ git checkout -b tighten-triage
$ kortix cr open --head tighten-triage --title "Tighten the triage prompt"
```

Source: https://kortix.com/docs/git

**20,000+** GitHub stars

## Open source. Read it, fork it, audit it.

[Kortix on GitHub](https://github.com/kortix-ai/suna) · [Kortix docs](https://kortix.com/docs)

_FAQ_

## Questions people ask before they switch.

Short answers about Kortix, the open-source option we recommend for the agent layer.

### Is Langflow open source?

Yes. Langflow is an open-source visual builder. Kortix is open source too, and the two solve different jobs: Langflow wires a flow you design, and Kortix runs agents that finish open-ended work.

### Is Kortix an alternative to Langflow?

Kortix fits when the job is open-ended work an agent completes, not a fixed sequence you draw. Langflow keeps the canvas; Kortix keeps the work in a git repo you own, and session work reaches `main` through a change request.

### How is Kortix different from a workflow builder like n8n?

A workflow builder runs fixed steps in a set order. Kortix runs agents on their own cloud computers, each session on its own branch, and returns finished work. Kortix is a poor fit when the job is a fixed flowchart.

### Can I self-host Kortix?

Yes. Run it on Kortix Cloud, in your VPC, or on your own on-prem network. Self-host is free. The instance needs the internet, because kortix self-host start pulls images from docker.io.

### Which models can I use?

Any model provider with your own keys. Or the ChatGPT plan you already pay for. Or sign in with your OpenCode Console account for OpenCode Zen and Go.

### What does Kortix cost?

Free is $0 with 200 credits per month and 1 project. The team plan is $40 per seat per month with 2,500 credits per month per seat, pooled. Self-host is free.

## Run your whole company from one repo you own.

Start with one job and grow from there.

[Try Kortix](https://kortix.com) · [See the comparison](https://langflowalternative.com/#compare)

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