Datalumo turns your own documents into a private AI assistant your users can search and chat with. It runs on your server, under your own model keys.
You point it at your content. Datalumo reads, indexes, and stores it, all on your server. Then you use it three ways:
- A widget you drop on any page with one line of HTML.
- An HTTP API for your own app.
- A command-line tool for your terminal and scripts.
Your documents stay on your infrastructure, and Datalumo never sees them. When someone searches or chats, the question and the matching passages are sent to the AI provider you chose, on your key, to write the answer. That is the only data that leaves your server, and you choose which provider handles it.
Bring your own model key: OpenAI, Anthropic, Mistral, Gemini, or xAI. You pay the provider directly, with no markup. Or run a local model (Ollama, vLLM, LM Studio, or anything OpenAI-compatible) so nothing leaves your server at all.
- A server with Docker. Your laptop works for trying it out.
- An API key from a supported provider.
- Ten minutes.
One command:
curl -fsSL https://get.datalumo.app/install.sh | bash
It sets up everything, HTTPS included, and prints your admin key at the end.
Next, follow Getting started.