--- title: FAQ — Cromus AI Workflow Intelligence description: Frequently asked questions about Cromus, WIaaS, Croms™, SKILL.md, pricing, and the pre-execution governance layer for AI workflows. canonical: https://cromus.ai/faq source_html: https://cromus.ai/faq --- # Frequently Asked Questions --- ## What is Cromus? Cromus is the first **Workflow Intelligence as a Service (WIaaS)** platform. It provides a pre-execution intelligence layer for AI workflows — scoring, validating, simulating cost, and optimizing workflows before a single token is spent. It requires no external LLM calls; all analysis is deterministic. --- ## What is WIaaS? Workflow Intelligence as a Service (WIaaS) is a new AI infrastructure category Cromus defines. It operates *before* execution — upstream of orchestration frameworks (LangGraph, CrewAI), runtime agent stacks (Anthropic, Google Agent Platform), and observability tools (LangSmith, Helicone). WIaaS answers whether a workflow should run, in its current shape, on its proposed model, before any compute is consumed. --- ## What are Croms™? A Crom is one unit of preventable AI workflow waste. Croms are computed deterministically from a compiled SKILL.md and the verified model registry across four dimensions: cost waste, latency waste, failure risk, and structural inefficiency. The same workflow always produces the same Croms score. [Learn more →](/croms) --- ## What is SKILL.md? SKILL.md is an open specification for portable, structured AI workflow capability files. A SKILL.md captures execution profiles, model requirements, token estimates, input/output contracts, and governance constraints. Cromus compiles SOPs into SKILL.md files and validates them for free at [/validator/skill](/validator/skill). --- ## Does Cromus make LLM calls? No. All scoring, cost simulation, and Croms™ computation is deterministic — a pure function of the compiled SKILL.md and the verified model registry. No LLM calls are made during analysis. The same input always produces the same output. --- ## What AI models does Cromus cover? The verified model registry covers **62+ models** across **11 providers**: OpenAI, Anthropic, Google, xAI (Grok), DeepSeek, Mistral, Meta (LLaMA), Cohere, Amazon (Nova), Microsoft (Phi), and Qwen. Prices and deprecation status are checked weekly. --- ## What frameworks can Cromus export to? Cromus exports portable skill packs for: LangGraph, CrewAI, AutoGen, LangChain, LlamaIndex, n8n, Ollama, Continue.dev, Open WebUI, Claude Code, and Replit AI. --- ## What is the MCP server? Cromus ships a hosted Streamable HTTP MCP server at `https://mcp.cromus.ai/mcp` with six deterministic tools: `score_skill`, `simulate_cost`, `validate_skill`, `validate_ethos`, `validate_memory`, and `reality_check_summary`. It makes zero LLM calls and operates under a no-retention contract. It is compatible with Claude Cowork, OpenAI Codex CLI, and any MCP-compatible client. [Cowork connector →](/cowork-connector) | [Codex plugin →](/codex-plugin) --- ## How does pricing work? Every new account gets a **7-day free trial** with full Builder access — no credit card required. After the trial, the account drops to the free Lite tier (5 workflows, 10 skills) unless upgraded. Plans start at $9.99/month (Lite) and range to $249/month (White-Label). [See all plans →](/pricing) --- ## What is the Founding Member plan? The Founding Member plan is $19/month, locked forever. It is a limited cohort (25 spots) for early adopters who want to own their AI workflows at a price that never changes. [See if you qualify →](/founding-partners) --- ## Is there a free tool I can use without signing up? Yes. The SKILL.md, ETHOS.md, and MEMORY.md validators are free and require no account. You can validate any spec file at `/validator/skill`, `/validator/ethos`, and `/validator/memory`. --- ## How does Cromus relate to runtime agent stacks? Cromus is upstream. Runtime stacks (Anthropic Claude Managed Agents, Google Agent Platform, AWS Bedrock Agents, Azure agent gateways) govern and execute workflows that are already deployed. Cromus produces the governed, costed, validated input they consume — before execution begins. It does not compete with runtime stacks. --- ## Related pages - [What are Croms™? →](/croms) - [Pricing →](/pricing) - [Pre-Execution Governance →](/pre-execution-governance) - [Validators →](/validator) - [Contact →](/contact)