Anything LLM
A local-first all-in-one AI application that combines document Q&A with AI agents.
Download Anything LLM
macOS package (.dmg), architecture unlisted — verify before installing
Other downloads (1)
- Download
AnythingLLMDesktop-Silicon.dmg
unknown · .dmg · 483.3 MB
Getting a "can't verify the developer" warning?
If you see a warning on first launch, go to System Settings > Privacy & Security and click "Open Anyway". Open source software usually isn't signed with an Apple developer certificate, which is normal.
Clicking "Download" takes you to a source confirmation page first; the download starts after confirmation.
About Anything LLM
Stop renting your intelligence. Own it with AnythingLLM. Everything you need for a powerful local-first agent experience
See incorrect information? Tell us
Risk notice
Anonymous telemetry is enabled by default, collecting non-content usage data such as installation type and model provider. It can be disabled in the application settings or via environment variables. The desktop installers are not officially signed by Apple or Microsoft, so macOS may show a security prompt on first launch and require manual approval.
Who it's for
- Individuals or teams that want to deploy AI capabilities locally and maintain data sovereignty.
- Users who need Q&A and retrieval based on private documents.
- Professionals who want to build AI Agent automation workflows without code.
- Developers who want to self-host and flexibly integrate multiple LLM providers.
- Users who care about privacy and prefer not to use cloud-only AI services.
Who it's not for
- Users who prefer zero-configuration, out-of-the-box online services (model configuration is still required on first use).
- Users who are highly sensitive to any usage data collection and are unwilling to manually disable telemetry.
- Teams that require official technical support and enterprise-grade SLA guarantees.
Install guide
Installation Guide
Windows
- Intel/AMD 64-bit: Download
AnythingLLMDesktop.exe(recommended by this site) - ARM 64-bit: Download
AnythingLLMDesktop-Arm64.exe
After downloading, double-click to run and follow the prompts to complete the installation.
macOS
- Apple Silicon (M1/M2/M3): Download
AnythingLLMDesktop-Silicon.dmg - Intel chips: Download
AnythingLLMDesktop.dmg(recommended by this site)
Open the DMG file and drag the AnythingLLM icon into the Applications folder.
If you see a "cannot open because it is from an unidentified developer" message on first launch, go to "System Settings → Privacy & Security", click "Open Anyway", or right-click the app icon and select "Open".
Linux
- x64: Download
AnythingLLMDesktop.AppImage - ARM 64: Download
AnythingLLMDesktop-Arm64.AppImage
Grant execute permission and run:
chmod +x AnythingLLMDesktop*.AppImage
./AnythingLLMDesktop.AppImageFirst Run
After startup, you need to configure a large language model (LLM) provider. You can choose local models (e.g., Ollama, LM Studio) or cloud APIs (e.g., OpenAI, Anthropic), and establish a connection in the settings.
FAQ
Is AnythingLLM free software?
Yes. AnythingLLM is open-sourced under the MIT license, and can be used, modified, and distributed for free. It also supports self-deployment.
Which operating systems are supported?
Officially provided desktop installers are available for Windows, macOS, and Linux, including versions for Intel/AMD and ARM architectures.
Is an internet connection required to use AnythingLLM?
It depends on the model you choose. If you use local models (e.g., Ollama, LM Studio), it can run completely offline; if you call cloud APIs (e.g., OpenAI), an internet connection is required.
How do I update to a new version?
Please watch the GitHub Releases page and download the latest installer for your platform to overwrite the existing installation. You can also regularly visit the official website's download center for updates.
Will my data be uploaded?
It runs locally by default, and documents and data are stored on your device or self-hosted server. However, the project enables anonymous telemetry by default, which only collects non-content usage events such as installation type and vector database type. You can disable it in the settings.