--- title: "Flowise with EU-hosted models · ChatOpenAI setup guide" description: "Point Flowise's ChatOpenAI Custom node at inference.eu with a custom Base Path and build chatflows and agents on open-weight models processed in the EU." canonical: "https://inference.eu/integrations/flowise/" language: "en" --- Automation · Flowise # Use Flowise with open-source models hosted in the EU Flowise is an open-source, low-code builder for chatbots, agents and RAG pipelines. Its ChatOpenAI nodes accept a custom base path, so your chatflows and agentflows can run on open-weight models at inference.eu, processed in European data centres. [Contact sales](https://inference.eu/contact/)[Setup](#setup) 1. 01API format: OpenAI 2. 02EU data centres 3. 03Covered by one DPA ## Set it up in a few minutes 1. 1 Create an API key in your inference.eu account. 2. 2 In Flowise, go to Credentials, add an OpenAI API credential named `inference.eu` and paste your key. 3. 3 In a chatflow, drag in the ChatOpenAI Custom node from Chat Models and connect the new credential. 4. 4 Enter the inference.eu model id, for example `kimi-k3`, then open Additional Parameters and set Base Path to the inference.eu base URL. 5. 5 Connect the node to a chain or agent and send a test message in the chat panel. yaml ``` # Credential: Credentials > Add Credential > OpenAI API Credential Name: inference.eu OpenAI Api Key: # Node: Chat Models > ChatOpenAI Custom Connect Credential: inference.eu Model Name: kimi-k3 Streaming: true # Additional Parameters Base Path: https://api.inference.eu/v1 ``` ## What Flowise is Flowise is an open-source visual builder for language model applications. You drag nodes onto a canvas to create chatflows, built on LangChain components, and agentflows with multiple agents, tools, conditions and human review. Finished flows can be embedded as a chat widget or called through an API. Flowise runs self-hosted with Node.js or Docker, or as a cloud service. Chat models are nodes on the canvas. Flowise documents custom base URLs for its ChatOpenAI nodes so they can talk to any OpenAI-compatible service, and the ChatOpenAI Custom node accepts any model name. Together they connect your flows to inference.eu. ## How the configuration works Create an OpenAI API credential that holds your inference.eu key. Flowise stores credentials encrypted and lets you reuse them across flows. Then add the ChatOpenAI Custom node, connect the credential and type the inference.eu model id in Model Name. Under Additional Parameters, set Base Path to the inference.eu base URL ending in `/v1`. The standard ChatOpenAI node also has a Base Path field, but its model list contains OpenAI's names, so ChatOpenAI Custom is the simpler choice. Agentflow nodes let you pick ChatOpenAI Custom as their model in the same way. ## Which models to choose For conversational chatbots and retrieval chains over documents, `gpt-oss-120b` gives quick answers and keeps the chat widget responsive. For tool agents and agentflows with several steps, choose `kimi-k3` or `glm-5.3`, which produce dependable tool calls and handle long context. RAG flows that answer from long documents benefit from `glm-5.3` or `kimi-k3`. For flows that generate or review code, use `qwen3-coder-480b-a35b-instruct`. Each chat model node is independent, so one agentflow can route with a fast model and answer with a stronger one, all on inference.eu. ## Troubleshooting A 401 error means the credential holds the wrong key. A 404 usually means Base Path is empty, missing `/v1` or has an extra path; when Base Path is empty, the node calls OpenAI with your inference.eu key and fails. "Model not found" means Model Name differs from the inference.eu id. If a tool agent answers without calling tools, choose a model with stronger tool calling. If streaming does not appear in the chat panel, check Streaming is enabled on the node. Self-hosted Flowise may block outbound hosts through its HTTP deny list; make sure the inference.eu host is not on it. ## Keeping data in the EU With Base Path set to inference.eu, the messages, retrieved documents and tool results your flows send to the model are processed in EU data centres and never used for training. We sign a data processing agreement under Article 28 GDPR, publish our subprocessor list and bill in euros. Flowise stores flows, credentials and chat history in its own database. Self-host it in the EU for full control, and set `DISABLE_FLOWISE_TELEMETRY=true` to turn off anonymous usage telemetry. Embeddings, vector stores and tools in your flows use their own providers, so choose EU-hosted options for them. ## Flowise Category Automation Maker FlowiseAI API format OpenAI Processing EU data centres Website [flowiseai.com](https://flowiseai.com/) Documentation [docs.flowiseai.com](https://docs.flowiseai.com/integrations/langchain/chat-models/azure-chatopenai) ## Flowise: common questions The questions buyers, developers and legal teams ask us most. Something missing? Ask us directly. Yes. Flowise documents custom base URLs for its ChatOpenAI nodes. Use the ChatOpenAI Custom node, connect an OpenAI API credential with your key, enter the model name and set Base Path under Additional Parameters. Use ChatOpenAI Custom. It accepts any model name, such as \`kimi-k3\` or \`gpt-oss-120b\`. The standard ChatOpenAI node also has a Base Path field, but its model list shows OpenAI's own names. Open Additional Parameters on the ChatOpenAI Custom node and enter the inference.eu base URL, ending in \`/v1\`, in the Base Path field. Without it, the node sends requests to OpenAI's default address. Kimi K3 and GLM-5.3 are strong defaults for tool agents and agentflows. gpt-oss-120b suits conversational chatbots and retrieval chains where quick answers matter, and Qwen3 Coder 480B fits flows that work with code. Yes. Agent and LLM nodes in agentflows let you choose ChatOpenAI Custom as the model, with the same credential, model name and Base Path. Each node can use its own inference.eu model. Either the credential holds the wrong key, or Base Path is empty and the node is calling OpenAI with your inference.eu key. Check both, then save the flow and send a new message. Yes. Enable Streaming on the ChatOpenAI Custom node and the chat panel and embedded widget show tokens as they are generated. inference.eu streams in the standard Chat Completions format the node expects. Self-hosted Flowise collects anonymous usage telemetry unless you set \`DISABLE\_FLOWISE\_TELEMETRY=true\`. Flows and chat history stay in your database. On Flowise Cloud, they are stored on the provider's infrastructure under its terms. This page covers chat models. Embedding nodes and vector stores in your flows have their own credentials and settings. Choose EU-hosted or self-hosted options so that the documents you index stay in Europe with your model requests. Check outbound network and proxy settings on the server. Self-hosted Flowise also applies an HTTP deny list to Base Path values; if your administrator has customised it, make sure the inference.eu host is allowed. [Ask a question](https://inference.eu/contact/) ## Related [All integrations](https://inference.eu/integrations/) - [n8nPoint n8n's OpenAI credential at inference.eu and run AI Agent and Chat Model workflows on open-weight models processed in EU data centres.Automation](https://inference.eu/integrations/n8n/) - [DifyAdd inference.eu to Dify with the OpenAI-API-compatible plugin and build chatbots, agents and workflows on open-weight models processed in the EU.Automation](https://inference.eu/integrations/dify/) - [LangChainConnect LangChain's ChatOpenAI to inference.eu with base\_url and run chains, agents and LangGraph workflows on open-weight models processed in the EU.Frameworks](https://inference.eu/integrations/langchain/) ## Move your AI workloads to EU infrastructure Tell us which models you use and roughly how much traffic you have. We reply within one working day with pricing and next steps. - Reply within one working day - Pricing for your volume - DPA ready to sign