AnythingLLM
AI chat app for your documents, available as a desktop app or a self-hosted Docker server
In our rankings
Scores on the self-hosted ai chat interfaces criteria
Evaluated Nov 5, 2026- Services to run ×3 5
5/5 means: One container or one package, with storage built in
One Docker image with a built-in LanceDB vector database, so no separate database service is needed.
3 quotes - Model providers ×3 5
5/5 means: As 4, with ten or more named providers listed in the documentation
More than 30 named providers, a generic OpenAI option and a built-in local model engine.
4 quotes - Users and sign-in ×2 3
3/5 means: Several users with roles or per-user permissions
The Docker version supports several users with permissions; single sign-on, LDAP and two-factor sign-in are not mentioned on the pages we captured.
3 quotes - Chat with your documents ×2 4
4/5 means: Built-in document search with knowledge bases or workspaces, no extra service needed
Built-in document chat with a default vector database and nine to choose from, but hybrid search or multiple extraction engines are not documented on the pages we captured.
4 quotes
Evidence (18 quotes)
- 1
“AnythingLLM by default uses a built-in vector database powered by LanceDB”
- 2
“docker pull mintplexlabs/anythingllm”
- 3
“Supports both amd64 and arm64 CPU architectures.”
- 4
“Works with all popular closed and open-source LLM providers.”
github.com/Mintplex-Labs/anything-llm · checked - 5
“Run local models on your CPU or GPU (NVIDIA via CUDA, AMD and Intel via Vulkan) with no extra setup”
github.com/Mintplex-Labs/anything-llm · checked - 6
“Ollama (chat models)”
github.com/Mintplex-Labs/anything-llm · checked - 7
“Any open-source llama.cpp compatible model - OpenAI - OpenAI (Generic) - Azure OpenAI - AWS Bedrock - Anthropic - NVIDIA NIM (chat models) - Google Gemini Pro - Ollama (chat models) - LM Studio (all models)”
github.com/Mintplex-Labs/anything-llm · checked - 9
“AnythingLLM Docker is both a single-user and multi-user application”
- 10
“Multi-user instance support and permissioning”
github.com/Mintplex-Labs/anything-llm · checked - 11
“If you dont need multi-user support - you should use AnythingLLM Desktop.”
- 12
“Multiple document type support (PDF, TXT, DOCX, etc)”
github.com/Mintplex-Labs/anything-llm · checked - 13
“Intuitive chat UI with drag-and-drop uploads and source citations.”
github.com/Mintplex-Labs/anything-llm · checked - 14
“LanceDB (default) - PGVector - Astra DB - Pinecone - Chroma & ChromaCloud - Weaviate - Qdrant - Milvus - Zilliz”
github.com/Mintplex-Labs/anything-llm · checked - 19
“You can also do this in-app by going to the sidebar > Privacy and disabling telemetry.”
github.com/Mintplex-Labs/anything-llm · checked - 15
“AnythingLLM by Mintplex Labs Inc contains a telemetry feature that collects anonymous usage information.”
github.com/Mintplex-Labs/anything-llm · checked - 16
“Set DISABLE_TELEMETRY in your server or docker .env settings to "true" to opt out of telemetry.”
github.com/Mintplex-Labs/anything-llm · checked - 18
“You should aim for at least 2GB of RAM.”
- 17
“AnythingLLM Desktop is a " single-player " application you can install on any Mac, Windows, or Linux operating system”