Launch an agent on the AgentLeash hardened runtime
Managed agents run on our fork of OpenClaw with PAT verification built into the kernel. Permissions can't be bypassed, disabled, or tampered with. Ideal for enterprise, compliance, and government deployments.
Kernel-level enforcement. Unlike the plugin approach, managed agents have permission checks hardwired below the plugin layer. The agent has zero ability to circumvent its token.
Agent Identity
Name and configure your managed agent.
Unique identifier. Lowercase, hyphens, numbers.
Runtime Tier
Choose the compute and feature tier for your agent.
Starter
$29
per agent / month
- 1 vCPU, 2GB RAM
- 10,000 verifications/day
- Single chain (Base or Algo)
- 60s cache TTL
- 7-day audit logs
Popular
Professional
$99
per agent / month
- 2 vCPU, 4GB RAM
- 100,000 verifications/day
- Multi-chain support
- Configurable cache
- 30-day audit logs
- Webhook notifications
Enterprise
Custom
contact sales
- Dedicated infrastructure
- Unlimited verifications
- Multi-chain + private chain
- Custom cache policies
- 1-year audit retention
- SLA + dedicated support
- SOC 2 / FedRAMP ready
AI Model
Select the language model that powers your agent. You can change this later.
Model API costs are billed separately based on usage. AgentLeash does not mark up model pricing.
Deployment Region
Choose where your agent runs. Closer = lower latency for verification.
Permission Token
Assign an existing token or create a new one during deployment.
If you create a new token, you'll configure permissions after the agent is provisioned.
Deploying your agent...
Provisioning compute instance
Installing AgentLeash runtime
Configuring kernel-level PAT verification
Minting permission token on-chain
Running verification health check
Agent ready
Configuration Summary
Agent
—
Tier
—
Model
—
Region
—
Token
New
Monthly
—
+ model API usage (billed separately)