Models and the interfaces your team talks to, inside your perimeter. Prompts and documents stay on your side of the tunnel.
Open WebUI
A chat interface over models you host, so prompts and the documents people paste into them never leave your network.
Replaces ChatGPT Team, Microsoft Copilot
Dify
A builder for AI apps, agents and retrieval workflows that runs next to your documents and your models instead of at a platform vendor.
Replaces OpenAI Assistants, Azure AI Studio
LibreChat
One chat interface over every model your organisation allows, local or hosted, with the keys, the history and the choice of provider held on your side.
Replaces ChatGPT Team, Poe
Ollama
A server that loads open-weight models on demand behind an OpenAI-compatible API, run on GPUs in Switzerland, the EU or your own datacentre so prompts never reach a model vendor.
Replaces OpenAI API, Azure OpenAI
vLLM
An OpenAI-compatible inference server for open-weight models, run on GPUs in Switzerland, the EU or your own datacentre, so prompts and answers never reach a model vendor.
Replaces OpenAI API, Azure OpenAI
Qdrant
A vector database for semantic search and RAG, run on your own hardware or in Switzerland and the EU, so the embeddings of your documents stay where the documents are.
Replaces Pinecone, Weaviate Cloud
Langfuse
Tracing, evaluation and prompt management for LLM applications, run in Switzerland, the EU or your own datacentre so the prompts and answers it records stay with you.
Replaces LangSmith, Helicone
Choosing among the AI apps
These apps come in layers, and most teams need one from each. The interface is what people use every day: Open WebUI and LibreChat are chat assistants with your own models behind them, and Dify builds assistants and agent workflows on top of your documents. The model server runs the model itself: Ollama for a quick start on modest hardware, vLLM when many people share a GPU and throughput matters. Qdrant stores the embeddings that let an assistant answer from your own files, and Langfuse traces what each prompt did, what it cost and where it went wrong.
If you are starting out, pick one interface and one model server. Add a vector store when the assistant has to answer from your documents, and tracing when more than one team relies on it. Read the licence on each page before you commit: Open WebUI and Dify add conditions to a permissive licence, which matter if you want to change their branding or offer them to other organisations.
Wherever they run, on your own GPUs or in a Pilae region, prompts and documents stay on machines you control. Private AI describes the whole setup as one project.
Need an app that is not listed?
We scope any open-source app: what it needs, how we run it, what it costs.