HivemindOS manual
MiroShark And Runtime Gateways
MiroShark is an optional rehearsal room: it simulates how many agents would react to a plan before you run it for real. This page also covers runtime gateways, which let runtimes such as OpenClaw act as agents in Fleet and Chat.
To install and connect MiroShark, see MiroShark setup.
What you can do with MiroShark
- Install, start, and open MiroShark from the dashboard.
- Start a simulation from the Swarm view, or type
/swarm-sim <scenario>in chat. - Pick a template, fill in its fields, and set rounds and platforms (X, Reddit, Polymarket-style markets, and more).
- Follow a run’s timeline, market events, and outputs, and read its summary, report, and transcripts.
- Load past runs from your shared vault.
- Ask a helper agent to suggest a template or rewrite a scenario.
- Have a dashboard agent analyze a run.
- Send results into your agents’ work or the Work Board.
- Publish supported runs.
HivemindOS keeps the real work — tasks, handoffs, and decisions. MiroShark only rehearses.
Hosted runs and cost
Local and self-hosted MiroShark is free.
You can also run a hosted simulation from the Swarm view or chat for $1.20 per run. You pay from your shared HivemindOS balance or directly from a selected crypto wallet through x402. The app remembers your payment choice but always shows the price and asks you to confirm before each run. If the run does not start, the user is not charged. Choosing a hosted run is the only action that creates the $1.20 cloud charge.
MiroShark as a fleet service
If MiroShark runs on another machine in your hive, My Apps can still list its API, grouped into templates, run lifecycle, run data, graph, observability, and config.
Runtime gateways
When you add an OpenClaw runtime as an agent, HivemindOS connects to it through a small gateway. That gives you:
- automatic detection of the runtime profile;
- model selection from the runtime’s local config;
- chat with it from the normal HivemindOS chat.