Self-host Flowise on a no-KYC VPS
Build LLM apps, chatbots, and agents with a visual drag-and-drop UI on your own server. Anonymous signup, no email required, no KYC. Pay with crypto or card, full root, no logs.
Quick start: Flowise on Servury via Docker
Tested on Ubuntu 24. Pick a 2 GB+ plan, deploy, SSH in.
# 1. Install Docker
curl -fsSL https://get.docker.com | sh
# 2. Run Flowise
docker volume create flowise_data
docker run -d --name flowise --restart unless-stopped \
-p 3000:3000 \
-v flowise_data:/root/.flowise \
-e FLOWISE_USERNAME=admin \
-e FLOWISE_PASSWORD=changeme \
flowiseai/flowise
# 3. Open http://YOUR_SERVER_IP:3000, log in
# 4. Drag nodes onto the canvas: LLM -> Vector Store -> Memory -> Output
# 5. Hit "Deploy" and Flowise gives you a REST endpoint + embeddable widget
What you build with Flowise
Drop a PDF/website node, a vector store, a chat model, done. Custom support bots that actually know your docs.
Compose LangChain agents visually: planner, researcher, writer, critic. Faster iteration than hand-writing chains in Python.
Every chatflow becomes a REST endpoint with auth. Drop it into your app, your iOS shortcut, your CRM, anything.
One snippet of HTML adds a chat widget to your site. Like Intercom but powered by your own LLM and your own data.
Connects to Ollama, LM Studio, vLLM, OpenRouter, OpenAI-compatible endpoints. Run fully offline if you want.
Chroma, Qdrant, Weaviate, Pinecone, Supabase, Postgres pgvector. Pick the one you already use.
Frequently asked questions
How much VPS do I need?
Flowise itself is light (~250 MB RAM). The reason 2 GB is recommended is to leave room for a vector DB, an embedding worker, and Postgres if you swap from SQLite. Heavy RAG workloads with millions of vectors deserve 4-8 GB.
Do I need an OpenAI key?
No. Flowise works with any OpenAI-compatible endpoint (Ollama, vLLM, LiteLLM, OpenRouter, Anthropic via wrapper, etc.). Run a local LLM and never send a token to OpenAI.
Can I version flows?
Flowise stores chatflows as JSON. Export them and commit to git. Some teams build a CI pipeline that imports flows on deploy.
Is it production-ready?
Used in production by thousands of small teams. For high-throughput public-facing chatbots, put a CDN/cache and rate-limit in front, monitor LLM costs, and pick a fast vector store.
What about Langflow vs Flowise?
Both visual LangChain builders with similar UX. Flowise is TypeScript/Node, Langflow is Python. Pick whichever lines up with your team's ecosystem. Both run fine on a Servury VPS.
How do I add custom tools?
Flowise supports custom JS tool nodes that you write inline. Anything from "fetch this URL" to "call my internal API with auth headers" works.
Can I auth the chat widget?
Yes, with API keys or JWT. Or put the widget behind your own auth gate and pass user identity into the prompt.
Does Servury log my Flowise traffic?
No. No application-level logging on customer servers. Anonymous signup, crypto/card payment, no logs.
