Open-source Latenode alternatives

n8n is a workflow automation platform for connecting apps, moving data between them and triggering actions without writing a full integration from scratch. Workflows are built on a visual canvas, and when a step needs more logic you can drop in JavaScript or Python and use npm packages. It is often compared with Zapier and Make, but the source is public and it can run on your own infrastructure.

Recent versions put a lot of weight on AI. You can build multi-step agents and LLM workflows that call OpenAI, Anthropic, Google or open-source models, add human approval steps, and inspect what happened afterward. The project advertises over a thousand integrations and a large library of ready-made workflow templates, and you can extend it with your own nodes.

n8n is distributed under a fair-code model: the Sustainable Use License, plus a separate enterprise license for additional features. The code is always visible and can be self-hosted, for example with Docker, but it is not an OSI-approved open-source license, so commercial redistribution is restricted. A managed cloud edition is available if you would rather not run it yourself.

Activepieces is an open-source replacement for Zapier, positioned as an all-in-one AI automation tool. Automations are built from pieces, which are type-safe TypeScript npm packages, and the interface is meant to work for both technical and non-technical users. The topics also describe it as an alternative to n8n. The repository metadata lists the license as Other.

Pieces that are contributed to the project automatically become available as MCP servers, so LLM clients such as Claude Desktop, Cursor and Windsurf can use them. The platform includes native AI pieces for working with different providers, an AI SDK for building agents inside the builder, and human-in-the-loop steps that delay execution or require approval. Developers can set up tools, then others in the organization use the no-code builder. It can be self-hosted, including in network-gapped environments, and supports branding customization.

Node-RED is a low-code programming tool for event-driven applications built on flow-based, visual programming. Instead of writing everything as code, you connect nodes that represent inputs, processing steps and outputs, which makes it popular for IoT projects, home automation and wiring together APIs and online services. It is a project of the OpenJS Foundation.

Getting started takes a couple of commands: install it globally with npm, start it with the node-red command, and open it in your browser. A large ecosystem sits around it, with a library of custom nodes and integrations, shared flows and node collections, plus documentation, a forum and a Slack community. Developers can also run the latest code from git by cloning the repository, installing dependencies and building it.

Node-RED is written in JavaScript on Node.js and licensed under Apache-2.0. You run it on your own computer, a Raspberry Pi, a server or in a container, so there is no hosted service to depend on. It suits makers, integrators and engineers who want to prototype and automate event-driven logic quickly.

Windmill is an open-source platform for developers to run internal code as APIs, background jobs, workflows and user interfaces. It is pitched as a self-hostable alternative to Retool, Pipedream and Superblocks and as a simplified version of Temporal, and it automatically generates UIs so a script can be shared and triggered as an internal app.

Scripts can be written in languages including Python, TypeScript, Go, Bash, SQL, PowerShell, Rust and GraphQL. They can be composed into flows or used within richer apps built with its low-code builder. The backend is written in Rust on PostgreSQL, and the README describes it as a very fast self-hosted workflow engine, citing a comparison with Airflow. Sections of its documentation cover main concepts, local development, architecture, security, performance and self-hosting with Docker Compose.

Windmill is open source under AGPLv3, and Windmill Labs sells dedicated instances, commercial support and licenses, so the repository license is listed as 'Other' on GitHub. It suits engineering teams that want to turn existing scripts into reliable internal tools and automations without building each interface by hand.

Pipedream is an integration platform for developers. It is a hosted platform, advertised as free for developers, for connecting applications and building event-driven automations. The GitHub repository holds the code for the pre-built integration components, the product roadmap and the documentation.

Workflows are sequences of steps, which can be pre-built actions or custom code, and each workflow is started by an event such as an HTTP request, a timer or a new row in a spreadsheet. Event sources emit events from services like GitHub, Slack, Airtable and RSS, while actions cover common operations such as sending a message or adding a spreadsheet row. More than 1,000 integrated applications are available, and any Node.js, Python, Go or Bash code can be run for custom logic.

The README says the company has demonstrated SOC 2 compliance. Because the platform itself is run by Pipedream, this is mostly a hosted service with an open repository of components rather than software you install, and the repository lists its license as Other. It suits developers who want webhook-driven glue code without managing servers.

ByteChef is an open-source platform that brings AI agents and workflow automation into one product. It can be self-hosted or embedded inside a SaaS application, and it is presented as an alternative to tools like Zapier and n8n, with a drag-and-drop workflow editor for building automations across APIs.

An AI Agent component runs the full agent loop of model, tool selection, execution and observation as a step in a durable workflow, so a failed or paused run resumes from that step. The README lists support for many LLM providers such as OpenAI, Anthropic, Azure OpenAI, Bedrock, Mistral, Groq, DeepSeek and Ollama, components that double as agent tools, several memory backends, and guardrails. Topics also mention MCP, embedded iPaaS and Workato-style integrations.

The core is licensed under Apache 2.0 according to the README, while the repository metadata lists the licence as other. ByteChef is written in Java and TypeScript, with an embedded edition for shipping the same agents in your own product, plus documentation, templates and a connector developer guide. It suits engineering teams that want automation and agents under their own control.