Anything LLM
AnythingLLM is a desktop and server application for chatting with your own documents, running AI agents, and swapping between language models without wiring up infrastructure by hand. You can run models locally, connect to providers like OpenAI or Azure, and work across PDFs, Word files, CSVs, codebases, and other formats you already store.
The product is built for people who want LLM workflows on their own hardware or inside their own environment. The desktop app installs on macOS, Windows, or Linux with no account required. Hosted cloud and self-hosted Docker deployments add multi-user access, tenant isolation, admin controls, and white-label branding for teams.
Privacy is the through-line. Desktop defaults keep models, embeddings, vector storage, chats, and documents on the machine running the app unless you choose otherwise. The project is open source under the MIT license, with a Community Hub for agent skills, system prompts, slash commands, and extensions.
One-click desktop install on macOS, Windows, and Linux with no account signup
Chat across PDFs, Word docs, CSVs, codebases, and imports from online sources
Run local models or plug in OpenAI, Azure, AWS, and other hosted LLM providers
Built-in agents, RAG, vector storage, and a developer API in one interface
Community Hub shares agent skills, system prompts, slash commands, and plugins
Cloud and self-hosted plans add multi-user tenants, admin controls, and white-label branding
Desktop app is free to download and runs fully local without creating an account.
Open-source MIT project with Docker self-hosting and a large GitHub community.
Supports local models plus OpenAI, Azure, AWS, and other enterprise LLM providers.
Bundles document chat, agents, RAG, vector storage, and a developer API together.
Cloud and self-hosted options add multi-user tenants, admin controls, and white-label branding.
Hosted cloud plans start at $50 per month, so managed team use is not free.
Basic cloud tier is limited to fewer than five users and under 100 documents.
Enterprise pricing requires a sales conversation rather than self-serve checkout.
Is AnythingLLM free to use?
Yes. AnythingLLM Desktop is free to download, and the open-source project can be self-hosted via Docker at no license cost. AnythingLLM also offers paid hosted cloud plans starting at $50 per month for teams that want a managed private instance.
Where are my documents saved in AnythingLLM?
On AnythingLLM Desktop, your documents, chats, models, and embeddings are stored locally on your machine by default. Cloud and self-hosted deployments keep data inside your private instance with tenant isolation between users.
What operating systems does AnythingLLM support?
AnythingLLM Desktop runs on macOS, Windows, and Linux. The site offers separate Apple Silicon and Intel downloads for Mac, plus installers for Windows and Linux.
Do I need an account to use AnythingLLM Desktop?
No. AnythingLLM Desktop is not SaaS and does not require signup to install or use the full local toolset. Hosted cloud plans use a separate checkout flow when you choose a managed instance.
How is AnythingLLM different from ChatGPT with a PDF plugin?
AnythingLLM runs on your desktop or your own server, lets you choose local or enterprise LLM providers, and keeps documents and chats under your control by default. It also bundles agents, RAG, vector storage, and a developer API in one app rather than relying on a single hosted chat interface.
What hosted plans does AnythingLLM offer?
AnythingLLM Cloud lists a Basic plan at $50 per month for smaller teams, a Pro plan at $99 per month with a 72-hour support SLA, and an Enterprise plan with on-premise installation and custom support. Self-hosted Docker is also available for free.

