AnythingLLM: chat over your document base
ChatGPT for your documents — ready in 5 minutes
- Video memory
- 8 GB
- Setup
- ~6 min
- Access
- port 3001
- Billing
- hourly, in BRL
You upload PDFs, pages and spreadsheets; it indexes them and answers citing the source. Because it runs on your machine and can use your own model, documents never pass through a third-party API — which settles much of the data-governance conversation.
AnythingLLM is full RAG with UI: ingests PDFs, websites, Notion, Confluence, GitHub. Connects to any LLM (OpenAI, Anthropic, Ollama, local vLLM). Multi-workspace, multi-user, agents, local embeddings, REST API. Open-source.
What it is for
- Manuals, standards and contracts searchable by anyone on the team
- Internal support answering from the official documentation
- Separate workspaces per team or per client
- You control the model, the logs and the prompt history
How to deploy
- Create your account and add balance (card or Pix, no subscription).
- In the console, pick the AnythingLLM template and a machine — the console hides the ones that do not meet the requirement.
- In about 6 minutes the setup finishes and the access address shows up in the panel, on port 3001.
Done? Just destroy the machine and billing stops with it. No contract, no minimum commitment.
FAQ
Do I need another template for the model?
Not necessarily: it can point at an external API. But the most common combo is this plus a vLLM or Ollama on the same machine.
How many users does it handle?
It is multi-user; the practical limit is the card serving the model.
Where do the documents live?
On the machine itself, in the template's storage folder.