Self-host AnythingLLM

AI Chat Interfaces

deployed & timed by usprices checked · Jul 2026

All-in-one self-hosted chat application with retrieval-augmented generation (RAG) over your own documents, organized into workspaces, and support for any local or hosted LLM and vector database.

ChatGPT / yr$240

Self-hosted / yr~$55

You keep$185/yr

Key facts

CategoryAI Chat Interfaces

LicenseMIT

StackJavaScript, Node.js, React

Min RAM2048 MB

Dockeryes

Difficulty

Our recommendation

Pick AnythingLLM when the point is chatting with your own documents — it's an all-in-one app that bundles RAG, workspaces, and a vector database, and connects to any local or hosted LLM. It's the middle path: more batteries-included than Open WebUI for document Q&A, simpler than assembling LibreChat's stack. If you don't need retrieval over your files, a plain chat UI is less to run.

What you need

Any VPS with at least 2048 MB of RAM
A domain you control — most self-hosted setups need HTTPS in front of them
Under an hour if you've used Docker before

Install with Docker Compose

Save this as compose.yml and run docker compose up -d:

# AnythingLLM — official Docker quick start
export STORAGE_LOCATION=$HOME/anythingllm
mkdir -p $STORAGE_LOCATION && touch "$STORAGE_LOCATION/.env"
docker run -d --rm -p 3001:3001 \
  --cap-add SYS_ADMIN \
  -v ${STORAGE_LOCATION}:/app/server/storage \
  -v ${STORAGE_LOCATION}/.env:/app/server/.env \
  -e STORAGE_DIR="/app/server/storage" \
  mintplexlabs/anythingllm  #  →  http://SERVER_IP:3001

vps picksBest VPS for self-hostingbenchmarked quarterly · re-priced monthly

What you take on

AnythingLLM does document RAG out of the box, but retrieval quality and the model behind it are still yours to own:

non-negotiableRAG is only as good as its setup. Chunking, embeddings, and the vector store all shape answer quality — expect to tune them, and to see confidently-wrong answers when retrieval misses the relevant passage.

non-negotiableIt still needs a model. AnythingLLM is the app around an LLM, not the LLM — point it at a local runtime (Ollama, LocalAI) or a hosted API, and that choice sets both your privacy and your hardware bill.

non-negotiableThe container wants SYS_ADMIN. The official quick-start adds the SYS_ADMIN capability for its document processing — understand what that grants before running it on a shared host.