Flowise Alternatives: Where to Move After the Shutdown

If you need Flowise alternatives, it's because Flowise is shutting down. The team announced the wind-down with a code freeze on July 29, 2026, archived the GitHub repository in August, and reached end of life on August 31, 2026. After that date the core team stopped supporting it on Discord and GitHub. This came a year after Workday acquired Flowise in August 2025. The code is Apache 2.0, so you can fork it, but nobody is maintaining it upstream anymore.

The right replacement depends on what you built in Flowise:

• Visual LLM and agent flows (Chatflow, Agentflow) → Langflow is the closest match
• A chatbot or assistant over your documents (Assistant, RAG) → Dify
• Agents that react to events in other systems → n8n
• Flows that outgrew the canvas → LangGraph or CrewAI in code
• A stable internal tool you can't move yet → keep running a fork while you plan

Quick comparison

Flowise's own docs describe three visual builders: Assistant for beginner-friendly chat agents with RAG over uploaded files, Chatflow for single-agent systems and simple LLM flows, and Agentflow, which covers both of those plus multi-agent systems and complex orchestration. Work out which one each of your flows used. That usually tells you where it should go.

If you built visual LLM and agent flows: Langflow

Langflow is the most direct replacement. You still drag components onto a canvas and wire them together. You still test in a chat panel (Langflow's is called the Playground). You still call the finished flow through an API. For anyone searching "langflow vs flowise," the question has changed: it's no longer which one to pick, it's what changes when you move.

The language changes. Flowise was a Node.js and TypeScript app, built on LangChain's JavaScript packages. Langflow's documentation calls it "an open-source, Python-based, customizable framework," and its custom components are written in Python. Any custom JavaScript you wrote in Flowise has to be rewritten, not ported.

You'll rebuild, not import. Langflow's import feature accepts Langflow JSON files. A Flowise export won't load. Plan to recreate each flow by hand, using your Flowise version as the spec.

MCP works in both directions. Langflow works as both an MCP client and an MCP server. As a client, its MCP Tools component lets a Langflow agent use tools from any MCP server. As a server, it exposes a project's flows as tools that MCP clients can call, over streamable HTTP with SSE as a fallback. If your Flowise flows were destined to become tools for Claude or another agent, this is the cleanest path.

Installing it looks different. The docs recommend Langflow Desktop, a standalone app for macOS 13 or later and Windows. Docker and a Python package (Python 3.10 to 3.14) are the alternatives. Langflow also runs a managed cloud, and its site says it's the same Langflow whether you use the open-source version or the cloud.

Check who owns it, since Flowise shows why that matters. Langflow is MIT-licensed. Its parent company, DataStax, agreed to be acquired by IBM in a deal IBM announced in February 2025. In that announcement IBM said it would keep supporting the open-source Langflow community. Flowise was also acquired, and it still shut down a year later. Being acquired tells you little either way. What does tell you something is whether commits are still landing. As of late September 2026, the Langflow repository is under active development.

Choose Langflow when:

• Your team still wants a visual canvas, not code
• Your flows are chatbots, RAG pipelines, or tool-using agents
• You want your flows to double as MCP tools
• Your team can support a Python runtime, or Langflow Desktop is enough

If you built a chatbot over your documents: Dify

Plenty of Flowise deployments were one thing: a chat assistant that answers questions from uploaded files, embedded on a website or an intranet. If that's you, the canvas was never the point. You need ingestion, retrieval, a chat interface, and hosting.

Dify bundles all four. Its documentation describes a platform for building agents, agentic workflows, and chatbots that draw on your own data, published as web apps or used through APIs. Much of what you assembled node by node in Flowise ships as one feature in Dify.

Two things to know before you move. Dify calls itself open source, but its licence is a modified Apache 2.0 with conditions on multi-tenant use and on keeping its branding. And it's a whole platform, which is a lot to run for one chatbot. Dify alternatives covers the licence terms and the other options in this category, so we won't repeat them here.

If your agents react to other systems: n8n

Some Flowise builds were really automations. A ticket arrives, an agent classifies it, and the result goes somewhere else. The LLM was one step in a larger process. Flowise could do this. An automation platform does it better.

n8n has an AI Agent node: you connect a chat model and one or more tools, and the agent decides which tools to call. Around that node sit triggers, schedules, error handling, and a large library of integrations with ordinary SaaS tools. That's the part Flowise never really had. n8n also connects agents to MCP servers.

n8n is written in TypeScript, which helps if your team wrote custom JavaScript in Flowise.

Know the commercial terms before you move. n8n is source-available under the Sustainable Use License, not OSI open source, and it limits use to internal business, personal, or non-commercial purposes. Per n8n's pricing, cloud plans are billed on monthly workflow executions, regardless of how many steps each one has. The Community Edition is self-hosted. If you're choosing between open automation tools, n8n vs Activepieces compares n8n with a rival whose core is MIT-licensed.

If your flows outgrew the canvas: LangGraph or CrewAI

The Flowise team's own sunset note explains the shutdown with a change in how people build. Developers increasingly hand complex tasks to coding agents, and rigid low-code workflows hit their limits as complexity grows. If your Agentflows had turned into branching, looping graphs you could barely read on screen, that's a sign they belong in code.

LangGraph calls itself a low-level orchestration framework and runtime for long-running, stateful agents. It gives you durable execution, streaming, and human-in-the-loop steps, and it lets you mix fixed steps with steps the LLM decides. It has both Python and JavaScript versions. That makes it one of the few code options where a Flowise team can keep its TypeScript habits.

CrewAI is a Python framework built around Flows, which orchestrate the work, and Crews, which are teams of agents working on delegated tasks. It fits multi-agent Agentflows where the work splits cleanly into roles.

Both mean trading a UI for a repository your team owns. LangChain alternatives goes deeper on choosing a framework and on when to skip one entirely.

If you can't move yet: keep running your fork

This is a legitimate option, and the Flowise team suggests it. The code outside the enterprise directory is Apache 2.0, and the announcement tells teams to fork the repository and maintain their own updates.

Be clear about what you're taking on:

A fork makes sense for an internal chatbot that isn't exposed to the internet, doesn't change much, and has an owner. It's a bridge, not a destination. Pin the version, back up your database and flow definitions today, and put a migration date on the calendar. Anything customer-facing or internet-exposed should move sooner, because security fixes are now your job.

The step before rebuilding

Every option above means rebuilding: new components, new wiring, rewritten prompts. For a flow your business depends on, that work is worth doing properly.

Sometimes, though, the flow was never really yours. You set it up in Flowise because someone shared a working setup, and what you actually want is to keep using it, not to rebuild it on a new platform. Taku fits that case. It's an AI-native desktop workspace where you mirror an AI app, skill, or workflow someone else already got working, run it on your own machine, and remix it into your own. It isn't a flow builder and doesn't compete with Langflow or n8n. It's for when starting from a working setup beats designing one. Taku is in Beta, and the Mac app is available now. The free app library shows what that looks like.

Picking among Flowise alternatives

1. Visual Chatflows or Agentflows you want to keep visual? → Langflow.
2. A RAG chatbot or assistant over your documents? → Dify.
3. Agents triggered by tickets, emails, or other systems? → n8n.
4. Flows too complex to follow on a canvas? → LangGraph (Python or JavaScript) or CrewAI (Python).
5. Want your flows to double as MCP tools? → Langflow.
6. Stable internal tool and no time right now? → fork, pin, back up, and schedule the move.

Whatever you pick, back up your flows today. Nobody upstream will do it for you.

FAQ

Is Flowise shut down?

Yes. The team announced the wind-down with a code freeze on July 29, 2026, archived the GitHub repository in August, and reached end of life on August 31, 2026. The code stays public on GitHub, but the core team no longer maintains it or offers support.

Didn't Workday buy Flowise?

It did. Workday announced the acquisition on August 14, 2025. The sunset announcement came about a year later and is signed by the Flowise team.

Can I keep using Flowise?

Yes. The code outside the enterprise directory is Apache 2.0, and the team encourages forking it. You just become responsible for security fixes and dependency updates. That's workable for a low-risk internal tool, and risky for anything exposed to the internet.

Langflow vs Flowise: which is closer?

Langflow is the closest replacement: the same visual-canvas approach, plus an API and MCP support. The big differences are that Langflow is Python where Flowise was TypeScript, and you'll need to rebuild your flows by hand because Langflow imports only its own JSON format.

Can I import Flowise flows into Langflow?

No. Langflow's import feature takes Langflow JSON files. Use your Flowise flows as the spec and rebuild each one.

Which Flowise alternative is fully open source?

Langflow, LangGraph, and CrewAI are MIT-licensed. Dify uses a modified Apache 2.0 licence with extra conditions, and n8n is source-available under its Sustainable Use License. Neither of those two is OSI open source.