Norma
Norma by Quality Clouds is a governance and rule enforcement service designed specifically for AI-generated code. It integrates directly into AI coding agents through the Model Context Protocol (MCP), supporting editors such as Cursor, Claude Code, Lovable, Replit, and any other MCP-capable client. Before code is committed, Norma checks it against an organization's rules and answers three critical questions: does it follow policy, are any secrets or credentials exposed, and is the architecture sound. This ensures that AI-assisted development stays aligned with enterprise standards and security requirements. Norma performs its checks outside the AI model, returning deterministic results that include a rule ID, severity level, line number, and suggested fix. This approach eliminates the inconsistency that can occur when asking a model to evaluate code multiple times. The same file, rules, and agent will always produce the same verdict. Norma also provides a centrally versioned ruleset, preventing standards from drifting across multiple repositories. Every violation and applied fix is recorded and tagged with the model and version, creating an auditable trail. Setup is designed to be fast: users can connect Norma to their agent in under 30 seconds with OAuth sign-in and a single configuration entry. The service is free to start permanently and audits every repository connected to it. Norma is positioned for individual builders, engineering teams, and enterprises that want to enforce governance on AI-generated code without slowing down development.
- Category
- Code
- Pricing
- Free
Problem solved
AI coding agents can produce code that violates organizational policies, exposes secrets, or contains architectural flaws. Models often give inconsistent answers when asked to review the same code. Norma solves this by enforcing rules deterministically outside the model, providing a repeatable verdict with rule ID, severity, line, and fix, and centralizing ruleset versioning across repositories.
Use cases
- Enforce coding standards in AI-generated code before commit
- Detect exposed secrets and credentials in agent-written code
- Audit repositories for policy compliance
- Standardize governance across multiple repositories
- Integrate rule checks into CI/CD via MCP
- Provide deterministic code review for AI agents
- Onboard new developers with automated policy checks
Audience
- Developers
- Enterprise
- Startups
- Hobbyists
Pros
- Deterministic, repeatable code checks outside the model
- Works with any MCP-capable AI editor
- Centralized, versioned rulesets prevent drift
- Free to start, no credit card required
- Quick setup with OAuth and one config entry
Cons
- Requires an MCP-capable editor or client
- Cloud-hosted service may require sending code to external servers
- Free tier may have limitations for large teams
- Rule library may need customization for specific org needs
Official website: https://qualityclouds.ai/norma
Canonical LambdaBase page: https://www.lambdabase.com/tools/norma