OpenClaw-first agent operations

Run OpenClaw agents in production
from one self-hosted control plane.

Lynxrun supervises isolated OpenClaw processes on one VPS with logs, alerts, backups, encrypted secrets, and cost visibility from one private dashboard.

Get started View Docs
curl -fsSL https://lynxrun.com/install.sh | sudo bash
Agents
Runtime: OpenClaw 3 running 1 starting 1 crashed
research-agent anthropic/claude-sonnet-4-20250514
1.9k $0.88
code-reviewer openai/gpt-4.1-mini
820 $0.56
email-dispatcher anthropic/claude-3-5-haiku-20241022
1.2k $0.12
ops-triage gemini/gemini-2.5-flash
640 $0.09
data-transform ollama/llama3.2
crashed $0.00

What it does

Operate OpenClaw agents
from one control plane.

Lynxrun is the operations layer around OpenClaw: it launches isolated processes, supervises them, streams logs, stores encrypted secrets, and keeps recovery and cost tools in one place.

  • Run each agent as an isolated OS process on one VPS
  • Supervise restarts and watch running, starting, stopped, crashed, and unhealthy states
  • Stream logs and alerts from one private dashboard
  • Store OpenClaw model-provider keys and channel secrets encrypted at rest
  • Download backups and stage restores without leaving the product
  • Track per-agent usage and estimated cost as workload grows
Isolated processes
Each agent runs in its own supervised OpenClaw process.
Cost visibility
See usage and spend by agent so expensive work is visible early.
Backups and restores
Download a backup or stage a restore for the next restart.
Encrypted secrets
Provider keys and channel configs are encrypted before persistence.

How it works

Install from the docs.
Then open the dashboard from your device.

01
Install on your server
Follow the getting-started guide to install Lynxrun on your VPS. The dashboard stays on 127.0.0.1:18789, not on a public port.
02
Open the private dashboard
Use the SSH command from the installer, or Tailscale Serve if you want another device to connect. Your first browser load opens the in-app setup.
03
Add a model provider and launch
Add an OpenClaw model provider, deploy your first agent, and start operating it from the same control plane.

Who it's for

Built for operators
running OpenClaw on one VPS.

  • Operators who want to run agents on their own VPS
  • Builders who want supervision, logs, alerts, and restarts without hand-managing processes
  • Self-hosters who want private access over SSH or Tailscale, not a managed cloud
  • Anyone who wants OpenClaw now and understands more runtime adapters come later

Why Lynxrun

Self-hosted
Keep your agents, credentials, and operations on infrastructure you control.
Operations layer for OpenClaw
Lynxrun handles supervision, logs, alerts, backups, and cost while OpenClaw remains the shipped runtime.
OpenClaw now, more adapters later
Launch ships with OpenClaw only. Future runtime adapters can come later without pretending they already exist.

Pricing

Start free. Pay only for higher agent limits.

Every plan keeps the same self-hosted product, private dashboard, OpenClaw model-provider setup, logs, alerts, backups, encrypted secrets, and cost visibility. Paid plans only raise the agent cap.

Community
Free
Install Lynxrun and run up to 3 agents on your own VPS.
  • Up to 3 agents
  • Install from docs, then finish setup inside your own instance
  • Includes logs, alerts, backups, encrypted secrets, and cost tracking
  • Uses the same private dashboard and onboarding flow as paid plans
  • Upgrade later without changing your setup
Start Free (Community)
Hobby
$ 249
For operators who need more room without changing the self-hosted workflow.
  • Up to 10 agents
  • Upgrade from Community without changing your setup
  • Keep the same dashboard, onboarding flow, and provider setup
  • Keep the same logs, alerts, backups, and cost visibility
  • Stay on the same single-VPS workflow as you add more agents
Buy
Team
$ 799
For installations that need unlimited agents on the same self-hosted product.
  • Unlimited agents
  • Keep the same dashboard, onboarding flow, and provider setup
  • Keep the same logs, alerts, backups, and cost visibility
  • Use the same Lynxrun instance and workflow
  • Scale agent count without changing product scope
Buy

Install Lynxrun and operate OpenClaw agents
from one VPS.

Follow the install guide, open the private dashboard, add an OpenClaw model provider, and launch your first agent. OpenClaw ships now. More runtime adapters can come later.

Get started