AnythingLLM: Review, Features and Pricing
anythingllm.com · checked 28 Sept 2026
A private AI assistant for documents, meetings and web tasks on desktop, mobile, cloud or your own server.
How it plugs in
Signal on Byok System6.7#3 of 24
AnythingLLM is a private AI assistant that can run on a computer without an account, API key or token limit. It builds document knowledge bases, handles meeting recording and summaries, and offers web search, file uploads and image generation. Desktop, mobile, cloud and self-hosted options are available, with free plans and paid cloud tiers.

Review
In a line
AnythingLLM is a private AI assistant for desktop and mobile use. It can also run as a cloud service or on your own servers.
What it does
The app turns your documents into an agent knowledge base while keeping the data on your device. It can search the web, accept file uploads and generate images.
Its Meeting Assistant records, transcribes, summarizes and indexes meetings on the computer. It does not require a bot to join the call or cloud processing.
The desktop app also includes web scraping, model selection, background jobs and custom agent skills. Mobile adds document chat, file generation, scheduled jobs and memory while running small models on the device.
How it connects
AnythingLLM connects with providers such as Ollama, LM Studio, OpenAI, Anthropic, Google Gemini Pro, Mistral API, Groq, Perplexity AI and others. It also supports app integrations and has a full Developer API for functions not already included.
The core software is provided under the MIT License. Self-hosted deployments keep documents, chats, workspace settings and embeddings on your infrastructure. Optional anonymous telemetry excludes document content and chat logs and can be turned off.
It runs on Android, web, Windows, macOS, Linux and through an API. Native desktop builds cover Apple Silicon and Intel Macs, plus Windows x64 or ARM.
Good fit
It suits people who want private document chat, meeting notes or web-assisted tasks. Teams can use cloud workspaces with multiple users, tenant separation and detailed administrator controls.
Self-hosting fits organizations that want deployment on their own servers. Cloud administration can also add white-labeling.
Watch for
Check Windows compatibility before installing: the documentation targets Windows 11 and a Windows Home user account, while other Windows versions may not work.
Cloud plans differ in administration and support. Enterprise pricing is custom, and the paid Basic plan requires you to bring an LLM API key. Choose a provider and deployment that match your privacy and management needs.
Pricing
Desktop and Self-hosted plans are free. Desktop has no account, API key or token limits; Self-hosted uses Docker for multi-user deployment on your servers.
Basic costs $50 a month, billed monthly, for a private instance, custom subdomain, RAG and agents, and requires you to bring an LLM API key. Pro costs $99 a month, billed monthly, adding priority resources and a 72-hour support SLA.
Enterprise is billed through Contact Us with price on request. It includes on-premise deployment, a custom SLA, custom integrations, SSO and RBAC.
Verdict
Choose AnythingLLM if keeping documents and meeting data close to your own devices or servers matters. Its range of model providers, deployment choices and free desktop and self-hosted plans make it flexible for individuals and teams.
Look elsewhere if your computer falls outside the supported Windows target, or if you need a simple fixed-price hosted service without choosing an LLM provider. Cloud buyers should compare the included administration and support before selecting Basic, Pro or Enterprise.
Our take
AnythingLLM is a practical choice for private AI work across documents, meetings and web searches. Free desktop and self-hosted editions lower the cost of trying it, while cloud plans add shared administration and support. The main decisions are where to run it, which model provider to connect, and whether your Windows setup is supported. It is less suitable for buyers who want a single hosted package with no provider setup, or who need predictable Enterprise pricing.
- Free desktop and self-hosted plans are available.
- Keeps self-hosted documents, chats and embeddings on your infrastructure.
- Supports many AI model providers, including OpenAI and Anthropic.
- Meeting Assistant records, transcribes and indexes meetings locally.
- Runs on desktop, mobile, web and self-hosted deployments.
- Windows documentation targets Windows 11 and a Home user account.
- Basic requires you to bring an LLM API key.
- Enterprise pricing is available only on request.
- Some cloud support and administration features are plan-dependent.
- Pick it ifPrivate document and meeting work; Teams needing self-hosted or multi-user AI; Users comparing many model providers
- Skip it ifYou need support for Windows versions outside the stated target.; You want one hosted price without arranging an LLM provider.; You require a published Enterprise price.
Pricing
Free Desktop and Self-hosted plans cover local or server deployment. Cloud Basic is $50 a month billed monthly, Pro is $99 a month billed monthly, and Enterprise is price on request.
- Basicmonthly · Private instance · custom subdomain · RAG · agents · bring an LLM API key$50/mo
- Promonthly · Private instance · RAG · agents · priority resources · 72-hour support SLA$99/mo
- DesktopNo account · no API key · no token limitsFree
- Self-hostedDocker deployment · multi-user deployment on your serversFree
- EnterpriseContact Us · On-premise deployment · custom SLA · custom integrations · SSO · RBACOn request
As anythingllm.com lists them, checked 28 Sept 2026.
Questions
Does AnythingLLM need an account or API key?
The free Desktop plan requires neither and has no token limits. Basic cloud requires you to bring an LLM API key.
Where does AnythingLLM keep self-hosted data?
For self-hosted deployments, documents, chats, workspace settings and embeddings remain on your infrastructure.
Which AI providers work with AnythingLLM?
Supported providers include Ollama, LM Studio, OpenAI, Azure OpenAI, Anthropic, Cohere, Google Gemini Pro, Hugging Face, Together AI, OpenRouter, Perplexity AI, Mistral API and Groq.
Can it handle meetings?
Its Meeting Assistant records, transcribes, summarizes and indexes meetings on the computer without a bot joining the call or cloud processing.
Is AnythingLLM open source?
The AnythingLLM core is provided under the MIT License.