MigrateFlows
Pipedream → n8n Agent Builder → n8n
Flowise → n8n migration

Move your Flowise chatflows and agentflows to n8n.

Flowise is archived and no longer gets security fixes. Drop your Flowise exports and get n8n workflows ready to import: models, memory, vector stores, tools, custom tool code and Agentflow v2 logic included.

—
days since Flowise reached end of life (August 31, 2026). The repository is archived; new vulnerabilities will not be patched.

Your Flowise flows

Drop your exports here Chatflow or Agentflow JSON exports, the "Export all" file from Flowise settings, or a ZIP with several of them.
  1. Export from Flowise. In each flow, open Settings and choose Export. To include your custom tools' code, use Settings → Export → All instead.
  2. Convert here. Each flow becomes an n8n workflow with AI Agent, chains, models, memory and vector stores wired the same way. RAG flows also get a document loading workflow.
  3. Import into n8n. Use Import from file. Add your credentials from the checklist, run the document loading workflow if there is one, and chat with it.

Everything runs in your browser. Your flows are not uploaded anywhere. Flowise exports never include API keys.

Converted for n8n

Your converted workflows will appear here.

Converts automatically

  • Chains and agents: LLM Chain, Conversation Chain, Retrieval QA, Tool Agent, ReAct and conversational agents, supervisor and worker teams
  • Chat models, embeddings, memory, text splitters and output parsers (zod schemas included)
  • Vector stores keep pointing to your Pinecone, Qdrant, Supabase, Postgres, Chroma, Milvus, Weaviate, Redis or MongoDB data
  • Custom tools: your JavaScript runs in n8n's Code Tool, including fetch and node-fetch calls
  • Agentflow v2: LLM, Agent, Condition, Condition Agent, HTTP, Custom Function, Iteration, Loop, Direct Reply and Human Input

Needs a quick check

  • In-memory and FAISS stores: run the generated loading workflow (or move to a hosted vector store)
  • Flowise Document Stores: their documents live in Flowise's database
  • Sequential agents (Agentflow v1) and some integration tools: rebuilt by hand or by us
  • MCP servers started by command (stdio): n8n connects to MCP over HTTP or SSE

Supported model providers

Done for you

Prefer to hand it off? We migrate your Flowise setup to n8n.

We convert every flow, rebuild what doesn't convert automatically, move your documents into a vector store n8n can use, set up credentials with you on a short call and test each bot with real questions before you switch.

Small

US$190

Up to 10 flows

Team

US$690

Up to 50 flows

Large

Custom quote

More than 50 flows, sequential agents or document stores

Email hello@migrateflows.com and tell us how many flows you have and where you want to run n8n.

No n8n account yet? Start with n8n

Questions

Can I keep running Flowise?

Your instance keeps working, but the project is archived: no new releases, no security patches and no support. Any new vulnerability stays open.

Do I need to re-upload my documents?

Not if they are in a hosted vector store such as Pinecone, Qdrant, Supabase or Postgres: the n8n workflow reads the same index, as long as you keep the same embeddings model. In-memory and FAISS stores are rebuilt with the loading workflow we generate.

Why n8n?

n8n has the same LangChain building blocks as Flowise (agents, chains, memory, vector stores, tools) plus hundreds of app integrations, and it is actively maintained, self-hosted or in the cloud.