Skip to content
ClawFleet logo

ClawFleet

Verified

Run isolated AI agent fleets locally via Docker and a browser dashboard.

Autonomous AgentsGeneral-Purpose 159Open source
View on GitHub
Updated 2026-06-15
ClawFleet GitHub repository

What is ClawFleet?

ClawFleet is an open-source platform for running teams of AI agents on personal hardware. It packages OpenClaw and Hermes runtimes into Docker containers that stay fully isolated from each other and the host system.

Users interact through a web dashboard that handles deployment, model connections, and instance control. Agents can be configured with reusable characters, skills, and memory snapshots while communicating via Discord or terminal access.

It targets developers and teams who want private, cost-free AI collaboration without relying on external SaaS platforms or exposing data to the cloud.

What you can build with ClawFleet

Local AI Team Building

Create multiple specialized agents with distinct personas that collaborate on tasks while keeping all conversations and data on your own machine.

Secure Experimentation

Test different models and agent configurations in sandboxed environments without risking host system access or incurring API overages.

Private Workflow Automation

Assign skills and tools to agents for ongoing internal processes such as research, coding assistance, or content planning entirely offline from cloud services.

Install ClawFleet

Install
curl -fsSL https://clawfleet.io/install.sh | sh
Quick start
curl -fsSL https://clawfleet.io/install.sh | sh
  1. 1Run the one-line install script to set up Docker, the CLI, and required images.
  2. 2Launch the dashboard daemon which opens at http://localhost:8080.
  3. 3Register your preferred LLM API keys in the Assets section.
  4. 4Create or select characters and assign them to new agent instances.
  5. 5Start containers and access them via the dashboard, shell command, or Discord integration.

ClawFleet: pros & cons

Pros

  • +Strong isolation through per-agent Docker containers
  • +No ongoing cloud fees when using personal subscriptions
  • +Central dashboard for fleet oversight and quick terminal access
  • +Version pinning prevents unexpected runtime breaks

Cons

  • Requires Docker and works only on macOS or Linux
  • Initial image download is large and resource-heavy per instance
  • Advanced features like the character system are limited to OpenClaw
Did you find this helpful?

Frequently asked questions

No, agents run locally once models and images are downloaded, though initial provider connections need network access.

User reviews

Verified reviews from the community shape this listing's rating.

Loading reviews…

Sign in to review

Promote ClawFleet

Add this badge to your website, or share the tool.

DFeatured on DhanasviClawFleet 0