Run Dozens of Local Agents and Models
Your hardware shouldn't limit how many AI agents you run. BubbleHub schedules workloads smartly across your machines and comes with specialized agents ready to deploy, with no custom build required.
The hub for local agents
Run private AI agents on your own machine.
BubbleHub puts the agent era on hardware you already own. Keep your data private, avoid per-token cloud bills, and run AI close to the files and tools that matter.
curl -fsSL https://bubblehub.ai/install.sh | bash
Security Built In, Not Bolted On
Every agent is automatically sandboxed before it starts. A compromised or runaway agent stays contained, so your data and the rest of your system remain protected.
Access Only With Your Approval
When an agent needs to reach one of your apps, you get a prompt to approve or deny. Nothing connects without your explicit consent, so you stay in control every time.
One-click deployment
Tested on local agents
Download
Install in one line.
Start running local agents on your own hardware and keep everyday AI work private, predictable, and cost-effective.
Linux
curl -fsSL https://bubblehub.ai/install.sh | bash
Windows PowerShell
irm https://bubblehub.ai/install.ps1 | iex
Docs
Docs for local agents in production.
Read install guides, quick starts, CLI reference, sandboxing details, and release notes generated from the BubbleHub source on GitHub.
Open BubbleHub docs$ bubble --help
$ bubble models list
$ bubblehub
$ bubble run --root-dir . --binary ./examples/basic_agent.py --memory 16G
Build with us
The next era of agents belongs on hardware people control.
Enterprise
Bring local agents to your team.
Tell us how your team wants to cut AI costs, protect private data, and make better use of the hardware you already own.