Browser-based agent I can open now
OpenClaw
Best default for first agents, Codex-style sessions, and browser-first work.
// Agent selector
Tell Kriy.AI what kind of work you want to run. We host the runtime, route managed Kriy.AI tokens, show deployment progress, and keep feedback attached to the next run.
Start here
Most first users should choose OpenClaw.Browser-based agent I can open now
Best default for first agents, Codex-style sessions, and browser-first work.
Longer multi-step work with delegation
Use when the job needs planning, repeated execution loops, or multiple agents.
Controlled secure runtime
Use on Builder when isolation and controlled runtime experiments matter most.
// Supported agents
Start with behavior. Kriy.AI handles the runtime behind it.
agent / openclaw
Best first agent for most users.
Use OpenClaw when you want a direct hosted agent UI you can open and operate immediately.
runtime.profile / browser-first
agent / hermes
For longer-running and delegated workflows.
Use Hermes when the job needs planning, delegation, repeated execution loops, or multiple agents.
runtime.profile / multi-agent
agent / nemoclaw
For secure Builder environments.
Use NemoClaw when isolation, controlled runtime experiments, and security testing matter most.
runtime.profile / secure-builder
// Pricing
Every paid agent includes 100,000 starter tokens and a hosted workspace.
$19.99/mo
FREE first week - $19.99/mo after trial
1 vCPU / 2 GB RAM
Learning, tutorials, and first agents.
Try your first agent on research summaries, outreach drafts, study plans, or small spreadsheet cleanup.
$36.99/mo
Monthly subscription
2 vCPU / 4 GB RAM
Regular individual workflows and development.
Run recurring workflows like job tracking, meeting prep, weekly research, email follow-ups, and project planning.
$72.99/mo
Monthly subscription
4 vCPU / 8 GB RAM / 160 GB SSD
Heavier agent workflows and secure runtime experiments.
Use higher-capacity runtimes for long-running jobs, secure experiments, multi-step workflows, and controlled agent testing.
Managed tokens are installed server-side into the runtime. Users see status and balance, not provider keys.
// Setup flow
The first run should feel concrete. Create an account, choose the agent, confirm the plan, and Kriy.AI sets up the workspace behind the scenes.
01
Verify email and start the agent creation flow.
verify.email
02
Choose OpenClaw, Hermes, or NemoClaw based on the work you want to run.
select.runtime
03
Every new paid agent includes 100,000 starter Kriy.AI tokens.
tokens.included
04
Kriy.AI provisions the runtime, connects tokens, shows progress, and gives you a ready hosted agent.
workspace.ready
setup.process
ACCOUNT ─── AGENT ─── PLAN ─── WORKSPACE │ │ │ │ verify select tokens ready email runtime included to open
create_account()
└── select_agent()
└── checkout()
└── provision_workspace()
├── runtime.ready
├── tokens.connected
└── open_agent()workspace.status
// Managed stack
Agent selection should stay simple. The infrastructure, AI access, and product feedback loop run behind the scenes.
We provision the VM, runtime image, hostname, health checks, and visible progress state.
Every paid agent starts with 100,000 free tokens. Provider keys stay behind the gateway.
User feedback and support requests stay connected to product operations instead of disappearing.
// Technology Partners
