What is Dify.AI?
Dify.AI is an open-source LLM app development platform with a visual workflow builder. You can create AI agents, chatbots, RAG pipelines, and complex multi-step workflows without writing much code. It supports custom model providers through an OpenAI-compatible API configuration, and with 1.8 billion tokens and 747,000 requests on OpenRouter, it is one of the most popular platforms for building AI applications.
Dify supports adding custom model providers in its admin settings. Add EUrouter as a provider, enter your API key and base URL, and every model call from your Dify apps routes through EU infrastructure. This applies to all Dify features: chatbots, workflow nodes, agent tools, and RAG pipelines. If you are self-hosting Dify in the EU and using EUrouter, your entire AI stack runs within European jurisdiction.
Quick to integrate
A few lines of code is all it takes. Swap your base URL and you are routed through EUrouter.
1// Settings > Model Providers > OpenAI-API-compatible
2{
3 "Model Provider Name": "EUrouter",
4 "API Key": "eur-...",
5 "API Endpoint URL": "https://api.eurouter.ai/v1",
6 "Model Name": "gpt-4o"
7}Get started in minutes
Follow these steps to connect your application to EUrouter.
Open Dify model settings
In your Dify instance, go to Settings > Model Providers. Click "Add Model Provider" and select "OpenAI-API-compatible".
Configure EUrouter as a provider
Enter the API Endpoint URL as https://api.eurouter.ai/v1 and paste your EUrouter API key. Add the models you want to use (gpt-4o, claude-3-5-sonnet, etc.).
Build your AI app
Create a new chatbot, workflow, or agent in Dify and select EUrouter models. All LLM calls now route through EU infrastructure.
Why use Dify.AI with EUrouter
Visual AI workflows through Europe
Dify's drag-and-drop workflow builder lets you create complex AI pipelines visually. With EUrouter as the model provider, every LLM node in your workflow routes through EU servers. Build chatbots, RAG pipelines, and multi-step agents that are GDPR-compliant by default.
- Every LLM call in your Dify workflows routes through EUrouter's EU servers
- Chatbot, agent, and workflow nodes all use EU infrastructure
- RAG pipelines with EU-hosted embeddings for end-to-end compliance
- No code changes needed to make existing Dify apps EU-compliant
All models in one platform
Through EUrouter, Dify gets access to GPT-4o, Claude, Mistral, and every other model on the platform. Configure them all under one provider. Your workflow nodes can use different models for different steps, all through one API key.
- Access all EUrouter models from Dify's model selector
- Use different models for different workflow nodes
- One API key covers all models for all your Dify apps
- Switch models without reconfiguring your workflows
Self-host for full control
Dify is open source and designed for self-hosting. Run Dify in the EU, connect it to EUrouter, and your entire AI application stack is within European jurisdiction. The data flow is clear: user input goes to your EU-hosted Dify, which calls EUrouter in the EU, which routes to the model provider.
- Self-host Dify on EU infrastructure for complete data sovereignty
- EUrouter as the model provider keeps LLM calls in Europe
- Combined with EU-hosted databases, your stack is fully GDPR-compliant
- Docker-based deployment makes it straightforward to run in any EU cloud