Cloud compute: $0.461/hour while awake, $0 when idle or parked. Models priced at or below maker's rate. Example: Claude Opus 5.5 with $0.461/hour compute, or $0 model usage if using Claude Pro/Max subscription. Prepaid credits start at $10. Card top-ups include Stripe fee (2.9% + $0.30).
Starter: $7.50/month (monthly) or $9/month (annual); Pro: $15.83/month (monthly) or $19/month (annual); Premium: $40.83/month (monthly) or $49/month (annual). Annual plans save 17%.
Key features
Browser-based agent access
Agent API for integration
Multi-channel deployment via Slack and other apps
Native Mac app
Cloud-based agent execution
Bring your own Claude or ChatGPT subscription
Usage-based pricing with per-second compute metering
Agent skills for reusable instructions
Agent marketplace with 64+ agents
Upload PDF, Markdown, and text documents to seed simulations
Convert source material into ontology and graph structure
Run parallel simulations with up to 1 million agents per run
Generate prediction reports with key risks and confidence signals
Deep interaction through ReportAgent or direct character interviews
Dual-surface simulation across Twitter and Reddit style channels
Follow-up questions in the same workspace for continued analysis
Build executable scenarios for public opinion, market, policy, and narrative use cases
What makes it different
Access any agent in one place without per-agent switching
$0 model costs when using your Claude Pro/Max or ChatGPT Plus/Pro subscription
Cloud agents that cost $0 while parked or idle
Reusable agent skills that inject domain expertise automatically at runtime
Multi-channel deployment from browser to Slack to Mac without code
Inspectable AI forecasts treated as exploratory hypotheses before operational decisions
Review simulations before upload with the Forecast Explorer gallery
Graph-native simulation that exposes actor incentives, narrative spread, and reaction chains
Templates and guides for recurring scenario types like crises and product launches
Keeps judgment with humans on scenario boundaries and high-stakes operating calls
AgentSky