Runwaize builds the operations and distribution layer for AI agents in the real world.

Our ecosystem has four pillars:

• Supervaize — the commercial SaaS control plane for operating AI agents at scale (missions, jobs, HITL, governance, observability).
• Supervaizer — the open-source controller/runtime components that connect agents to Supervaize and enforce operational rules.
• Skills Cookbook — the skills catalog + curation + sharing layer (import skills, scan them, personalize them, publish libraries).
• Skills Cookbook Relay — the open-source local bridge that securely delivers curated skills to agents (MCP server, verification, local secrets).

We’re focused on one thing: making agent capabilities deployable, trustworthy, and manageable—without turning your agents into credential dumpster fires.

A user-first skills management layer that helps you turn “random skills from the internet” into a trusted, organized, shareable library for you and for your organization.

Core capabilities:

• Import skills (Vercel skills.sh, GitHub, uploads)
• Rate, comment, tag, and group skills into libraries
• Run security checks and attach scan reports
• Clone/personalize/version skills (configuration overlays by default, not forks)
• Manage variables (secrets as references by default)
• Publish libraries to third parties with install/update instructions

A lightweight local application that connects to Skills Cookbook and makes curated skills available to agents securely.

Core capabilities:

• Runs locally (desktop) and exposes a localhost MCP server
• Fetches latest approved skills/libraries from Skills Cookbook
• Verifies skill artifacts (hash/signature) before serving them to the agent
• Resolves variables and secrets locally (env/keychain/vault references)
• Caches safely and supports updates + revocations
• Keeps Skills Cookbook API credentials out of the agent runtime

The control plane for AI agent operations:

• Workspaces, missions, jobs
• Human-in-the-loop steps and approvals
• Visibility into agent behavior + outcomes
• Governance and policy scaffolding

Open-source controller/runtime components:

• Connect agents to operational tooling
• Collect telemetry and execution traces
• Enforce operational constraints (policy hooks, guardrails) where applicable
• Provide a foundation for integrations with external agents and frameworks