MA|FE
VibeCopGradusWhy Maiife?EnterpriseBook a Call
qwen3-32bllama-70bmistral-lgCOLLECTTRAINDEPLOYIMPROVE

VibeCop · a Maiife product

Drift becomes doctrine, catch it before it compounds.

VibeCop is Architecture Integrity & Product Intelligence for AI-assisted codebases. Every repository gets a 0–100 Architecture Integrity Index; every pull request is reviewed for architectural entropy — pattern drift, semantic duplication, hallucinated abstractions — before it merges.

Plan Review MCP works with Claude Code · Cursor · Codex · any MCP-capable agent

Works with

One source of truth. Every surface your team already uses.

Claude Code
Cursor
Codex
Copilot
Claude Desktop
Cowork
Windsurf
Continue
GitHub
Linear
Slack
Notion
Jira
Confluence
Anthropic
OpenAI
Gemini
Cohere
Mistral
Ollama
vLLM
LM Studio
200+ MCP Servers
Claude Code
Cursor
Codex
Copilot
Claude Desktop
Cowork
Windsurf
Continue
GitHub
Linear
Slack
Notion
Jira
Confluence
Anthropic
OpenAI
Gemini
Cohere
Mistral
Ollama
vLLM
LM Studio
200+ MCP Servers

Security & trust

Privacy-first. Enterprise-ready when you need it.

Defaults already pass review. Deeper governance lives on Enterprise — opt-in, same platform.

Content encrypted at rest

Ask-bar messages, knowledge-base documents and code-translation bodies use AES-256-GCM. Writes fail closed — with no key we refuse to store rather than store in the clear.

Data stays in Singapore

All data is stored in ap-southeast-1. Per-workspace region selection is not offered today.

Audit trail on governance changes

Policy, vendor, kill-switch and approval changes are recorded with before/after state and exportable to CSV.

Individual insights are private by default

A lead sees a member's metrics de-identified until that member opts in, per team.

Work-account sign-in

Google Workspace and GitHub today. Microsoft Entra ID and SAML are in progress for enterprise.

Policy-checked before the model

Every gateway request is authenticated, and proxied completions pass kill-switch, model-access, content-filter, budget, rate-limit and data-routing checks before reaching a vendor.

FAQ

VibeCop, explained.

What is VibeCop?

VibeCop is Maiife's Architecture Integrity & Product Intelligence platform. It reviews AI-generated code for architectural entropy — hallucinated abstractions, semantic duplication, pattern drift — and scores every repository with a 0–100 Architecture Integrity Index, so structural problems surface while they are still cheap to fix.

What is the Architecture Integrity Index?

The Architecture Integrity Index is a 0–100 score of a codebase's structural health, computed as 100 minus entropy impact. It weighs four axes: Pattern Coherence, Knowledge Reuse, Abstraction Discipline, and Architectural Complexity.

What does VibeCop detect?

VibeCop runs ten architectural detector agents across four axes — hallucinated abstractions, semantic duplication, over-engineering, pattern drift, reinvented utilities, architectural complexity, dependency health, layer violations, orphaned modules, and reintroduced bugs — plus six static hygiene layers (dependency CVEs, secrets, SAST, infrastructure-as-code, vendored drift, magic numbers) tracked separately from the index.

Can VibeCop review my AI agent's plan before code is written?

Yes. The VibeCop Plan Review MCP is a remote MCP server that checks an AI coding agent's implementation plan against your codebase's conventions before any code is written. It works with Claude Code, Cursor, Codex, and any MCP-capable agent.

Does VibeCop work with GitHub?

Yes. VibeCop connects to GitHub repositories and reviews pull requests automatically. Webhook-triggered auto-reviews on every PR are available on paid plans.

How is VibeCop different from a linter?

Linters check line-level style and syntax. VibeCop analyzes architecture-level entropy — whether new code drifts from your team's established patterns, duplicates existing logic, or adds unjustified abstraction — using a full code graph of the repository.

Get in Touch

Your AI infrastructure needs a control plane.

Whether you're exploring our open source toolkit or ready for the full platform — we'd love to hear from you.