Olares
Your AI Agent. Your Data. Your Hardware.
Install Olares ·
Manage Olares with AI ·
Contribute
Olares is an open-source personal cloud OS you operate in plain language, built to run AI agents and LLMs on hardware you own.
Powered by Kubernetes, it turns your machines into a self-hosted AI platform accessible from any browser, giving everyone from individual users to small teams a unified place for compute, storage, networking, and apps.
https://github.com/user-attachments/assets/3e4b0c5d-dcbe-47eb-8628-88b062411fa0
🌟 If Olares is useful to you, consider giving the project a star. Your support encourages us to keep improving it.
Release highlights
- Olares 1.12.7: Router, Lares, and a broader local AI model catalog.
- Olares 1.12.6: Agent Skills, Model Console for local LLMs, and more reliable shared apps.
Why Olares
Great AI needs to know you. That requires access to your files, messages, and history. Many cloud AI services store this sensitive data on third-party servers and charge based on usage.
Olares brings AI home, so you can use Lares, the official Olares AI assistant, or agents like OpenClaw with local LLMs on hardware you own while still enjoying the access and convenience of the cloud.
Features include:
- One-click local AI: Install open-source AI apps and models from Olares Market with one click.
- Accelerated computing management: Pool GPUs and other accelerators across nodes, with time-slicing, memory-slicing, and exclusive GPU modes for AI, media, and gaming workloads.
- Files and storage management: Access local files, synced data, connected cloud storage, and external SMB/NFS shares through the built-in Files app, with configurable backups.
- Private networking and access controls: Use a private VPN, reverse proxy, and public, private, or internal entrances to give apps HTTPS endpoints without manually exposing individual ports.
- Anytime, anywhere access: Access your apps and files from your phone, desktop, or browser.
- A suite of system apps: Files, Vault, Market, Dashboard, Control Hub, and more, ready the moment you log in.
Get started
Linux script requirements
To install Olares on an existing Linux system using the script below, your machine needs:
- CPU: At least 4 cores
- RAM: At least 8 GB of available memory
- Storage: At least 150 GB of available SSD storage; installation will fail on an HDD
- OS: Ubuntu 22.04 to 25.04, or Debian 12 or 13
A dedicated GPU is optional and enables local AI acceleration.
Install on Linux
On your Ubuntu or Debian host, run the following command from an account with sudo access:
curl -fsSL https://olares.sh | bash -
This command downloads the official installer from olares.sh and runs it with Bash. See the Linux script installation guide for the full setup steps and troubleshooting.
For a fresh installation on an Intel or AMD x86-64 machine, use the ISO image. For a containerized installation on Linux, see the Docker Compose guide.
Once setup is complete, open your Olares desktop in a browser at its address, for example, https://desktop.marvin123.olares.com.
Try your first AI task
Follow the AI onboarding guide to set up Router, Lares, and a local model, then start your first conversation. Try asking Lares: “Check this device’s configuration.”
Key use cases
- Use a personal AI agent. Delegate research, file management, and routine automation to Lares through plain language, or use your preferred agent for coding and other tasks.
- Run generative AI locally. Chat with open models, generate images and video, and connect local models to other apps, all from your own hardware.
- Manage smart home devices and media. Connect home automation tools and stream your personal music and video libraries.
- Build and host agentic apps. Develop, test, and run apps and workflows in isolated environments on Olares.
- Process audio locally. Transcribe meetings, translate recordings, and generate speech without uploading audio to a third-party service.
- Create a self-hosted workspace. Give your family or team tools for documents, automation, project management, and communication.
- Manage your personal data. Store, sync, back up, and access files, photos, and documents across your devices.
System architecture
Just as public clouds offer IaaS, PaaS, and SaaS layers, Olares provides open-source alternatives to each of these layers.
For a detailed description of each component, refer to Olares architecture.
🔍 How is Olares different from traditional NAS?
Olares focuses on building an all-in-one self-hosted personal cloud experience. Its core features and target users differ significantly from Network Attached Storage (NAS) systems, which primarily focus on network storage. For more details, see Compare Olares and NAS.
Project navigation
The main directories in the Olares repository:
Olares/
├── apps/ # Built-in Olares system applications
├── cli/ # olares-cli: agent-native CLI with Agent Skills to install and operate Olares
├── daemon/ # olaresd, the system daemon process
├── docs/ # Project documentation
├── framework/ # Olares system services
├── infrastructure/ # Computing, storage, networking, and GPU components
├── platform/ # Cloud-native components like databases and message queues
└── vendor/ # Hardware-specific code for Olares devices
Contributing
Olares welcomes contributions across the project. Choose the path that best matches what you want to improve:
Learn more
- Installation guide: Choose an installation method and activate Olares.
- Use cases: Explore local AI, media, productivity, and self-hosted workflows.
- CLI guide: Install, manage, and diagnose Olares from the command line.
- Agent Skills: Let AI agents operate Olares through
olares-cli.
- Discord: Get community support and discuss deployments and agent workflows.
- Olares Forum: Share product feedback and join longer-form discussions.
- Follow Olares on X and YouTube.
Acknowledgements
The Olares project has incorporated numerous third-party open-source projects, including: Kubernetes, Kubesphere, Padloc, K3S, JuiceFS, MinIO, Envoy, Authelia, Infisical, Dify, Seafile, HeadScale, Tailscale, Redis Operator, Nitro, RSSHub, predixy, nvshare, LangChain, Quasar, TrustWallet, Restic, ZincSearch, filebrowser, lego, Velero, s3rver, Citusdata.
评论