AnythingLLM VPS hosting.

Yes. AnythingLLM is open-source software and free to self-host. On a ZAP-Hosting VPS you only pay for the VPS package you choose; current prices are shown live in the shop and depend on the VPS package (RAM, CPU and storage), term and location. If you connect a cloud LLM such as OpenAI or Anthropic, that provider bills its API usage separately. A local Ollama model adds no API cost.

• No license fee for AnythingLLM itself
• VPS price depends on package, term and location
• Cancel anytime

Order a VPS, open One-Click Apps in the dashboard and pick AnythingLLM. Coolify pulls the official Docker image, sets up persistent storage and starts the container, so you skip the manual Docker and Docker Compose setup. When the deployment finishes, open AnythingLLM in your browser and complete the setup wizard. Read the One-Click Apps guide

• 1. Choose a VPS package
• 2. Install AnythingLLM via One-Click Apps (Coolify)
• 3. Connect your LLM with an API key or Ollama

That depends on where the model runs. If AnythingLLM only talks to a cloud LLM API, the VPS mainly handles the interface, document parsing and embeddings, so a smaller package is often enough. Large document sets need more RAM while embedding, and a local Ollama model needs the most. Check the official AnythingLLM system requirements and move to a larger package later if your usage grows.

• Cloud LLM API only: smallest footprint
• Many large documents: more RAM for embedding
• Local Ollama models: the most RAM and CPU

No GPU is needed with a cloud LLM API: the model runs on the provider's servers and your VPS only sends requests. AnythingLLM supports providers such as OpenAI, Anthropic, Google Gemini, Mistral, Groq and OpenRouter, plus local Ollama or any OpenAI-compatible endpoint. Small local models can run CPU-only through Ollama, but replies are slower than with a cloud API.

• Cloud API: fast answers without a GPU
• Ollama on CPU: full data control, slower replies
• A different model per workspace

Every VPS includes a daily automatic Backup. With the ZapX Pass you also get manual Backups, snapshots and restore management. To update AnythingLLM, redeploy the service in Coolify to pull the latest image; your data stays in its persistent volume. For your own domain, point a DNS record to the VPS IP and add the domain in Coolify, which sets up HTTPS for it. Learn more about Backups

• Daily automatic Backup included
• Snapshots and restore with the ZapX Pass
• HTTPS for your domain through Coolify

Documents, embeddings, chat history and settings are stored on your VPS, not in a shared SaaS account. With a local Ollama model nothing leaves the server; with a cloud LLM only the prompt and the retrieved passages go to that provider. Multi-user mode adds password logins and roles, and with full root access you control SSH, the firewall and HTTPS yourself.

• Your data stays on your own VPS
• Admin, Manager and Default roles
• DDoS protection included

Yes. Switch on multi-user mode and invite your colleagues, each with their own login, role and workspaces. AnythingLLM Desktop runs on one computer for one person. The self-hosted version on a VPS stays online 24/7 and is reachable in any browser, so your whole team works with the same knowledge base from anywhere.

• Separate workspaces per team or project
• Browser access from any device
• One shared document base

ZAP-Hosting offers VPS locations in Europe and America. Pick the one closest to the people who use your AnythingLLM workspace, because that keeps the interface responsive for them. Many North American teams start with Ashburn, Dallas or Los Angeles; you see every available location when you configure your VPS.

• Locations in Europe and America
• Choose the region closest to your team
• Location is selected during configuration