How to deploy Flowise on Saplo: visual AI agent and chatbot builder without coding

Building your own chatbot or AI agent usually means one of two things: either writing integrations with a language model API from scratch, or a paid subscription to a closed SaaS platform through which all your customer data passes. Flowise is a third way: an open, visual AI agent and chatbot builder where you connect your LLM, custom tools (API calls, functions) and vector databases for document retrieval (RAG) in a drag-and-drop node graph, and then expose the finished flow as an API endpoint or a chat widget to embed on your site. In this guide, we'll show you how to set up Flowise on Saplo.

What is Flowise

Flowise works like a canvas with nodes: you add a language model node (OpenAI, Anthropic, a local model via Ollama, and others), connect tool nodes to it, such as an external API call or your own function, and optionally a vector database with your documents so that the agent answers based on company knowledge, not just the general knowledge of the model. You connect everything with arrows without writing integration code, test it in the panel, and publish it as an API endpoint or a ready-made chat widget to embed on a client's website. This is a different niche than n8n, which is a general automation engine without designing agent logic in the first place, and different from Open WebUI, which provides only a chat interface for a ready-made model but does not allow you to build your own agent logic with tools and memory. You can also find a description of the service on the Flowise hosting page on Saplo.

Who Flowise is for

• An agency or developer building AI features for clients - a chatbot, assistant, or document search engine, without writing API integrations from scratch for each project.
• A company that wants its own chatbot that answers based on company documents - without handing over their content to an external SaaS tool, RAG data stays on your Box.
• Anyone who tests and compares different LLM models and tools in one place - before deploying a specific solution to production.

Requirements

To set up Flowise on Saplo, you need:

• A Box with Docker running (the engine automatically enables it when installing from the catalog; container nesting is required, which Saplo has by default).
• A minimum of 2 GB of free RAM on the Box and 8 GB of disk space. The official Flowise image includes Chromium for web page and PDF scraping nodes, making it heavier than a typical single-container application. It does not fit with the Box operating system in S or M plans. The cheapest plan that can handle this is Box M for $68 per year.
• A Domain or subdomain *.saploapp.pl, included in the price of each plan, where the Flowise panel will be hosted.
• Your own API key for a language model provider (OpenAI, Anthropic, or other), which you will provide when building a flow in the panel, not during the installation itself.

For installation, you do not need any external database. Flowise by default stores flows and login data in a local SQLite file within the container's data volume.

Step-by-step installation

1. In the Saplo panel, go to App catalog and select Flowise (category "Business").
2. Select Box M or higher with at least 2 GB of free RAM, where Flowise will be installed, and the domain or subdomain under which the panel will operate.
3. Confirm the installation. The engine deploys a Docker container with the official flowiseai/flowise image, pinned to 3.1.3, exposes it only to the Box's loopback, and connects it to nginx, which serves it under the specified domain with free SSL. Only port 443 is exposed externally, not the raw container port.
4. The engine waits for the container to actually start, checking this after docker compose up completes, as the success of the command itself does not mean that the process inside did not crash a moment later.
5. After the installation is complete, the Flowise panel is available under the specified domain.

First-time setup

Flowise, like n8n, has a built-in account system and doesn't automatically generate access data during installation. Instead, it prompts you for them when you first access the panel.

1. Go to the domain address you specified during installation.
2. The first time you access Flowise, it will show you an account creation screen: enter your email address and password. Flowise generates its own session secrets on first launch and saves them in the data volume, so restarting the container doesn't log out all users.
3. After logging in, you'll see a list of flows. Click Add New, select Chatflow for a simple bot or Agentflow for an agent with multiple steps and tools.
4. On the blank canvas, add a language model node, for example, ChatOpenAI or ChatAnthropic, and plug in your own API key for the selected provider in its settings.
5. Optionally, add a tool node (custom API call, function) or a vector database node with your documents if you want the agent to answer based on company knowledge (RAG).
6. Connect the nodes with arrows, test the flow in the chat window on the right side, and when it works, publish it as an API endpoint or a chat widget to embed on a client's website from the Embed tab.

FAQ

Does Flowise require an external database? No, by default, it stores flows and accounts in a local SQLite file within the container's data volume. You don't need to set up a separate database.

What language models can I use? Any provider compatible with Flowise nodes: OpenAI, Anthropic, as well as local models via Ollama. In each case, you need your own API key or a locally running model; Flowise does not provide keys itself.

Why does Flowise require Box M and not Box S? The official Docker image includes Chromium for web page and PDF scraping nodes, making it significantly larger than a typical single-container application and requiring more RAM than the Box S plan provides, including the Box operating system.

How does Flowise differ from n8n available in the Saplo catalog? n8n is a general automation engine where AI agent logic is one of many possibilities, not the starting point. Flowise is designed around an LLM model, tools, and vector databases from the very first node, making it a better fit when you are specifically building an agent or chatbot, rather than a general integration flow between applications.

Is data in Flowise automatically backed up? It depends on your Box plan. Backup works at the Box level, just like with other applications from the catalog; it's not a separate Flowise feature.

If you have questions about setting up Flowise on your own Box or building your first AI agent or chatbot, write to us, and we'll help with the initial configuration. You can find Box plans and hosting prices on the Flowise on Saplo page.