Continuous cross-platform governance for AI coding agents

You use multiple AI coding tools.
Their configs drift.
Nerviq detects and aligns that.

Nerviq checks whether the AI tools in your repo agree on how to work. It finds drift, scores the current setup, and helps you fix the basics with reviewable changes.

Start with three ideas only: find drift, score the repo, and fix safely. The deeper surfaces can wait until after first value.

Find driftScore current repoFix safely
2,441
Repo checks
8
Supported platforms
3
Steps to first value
5
Public proof surfaces

The first problem Nerviq solves

Most teams do not have a single-tool problem. They have a repo-level coordination problem. Nerviq starts by telling you whether your active AI surfaces are actually aligned.

Contradictory instructions

One surface says run verification before every merge. Another never mentions it. Teams assume they are aligned until the repo starts drifting.

Silent ignore failure modes

A config file can exist, look correct, and still be ignored by the active agent surface. False confidence is worse than an obvious missing file.

Local setup, global inconsistency

Each tool can be locally coherent while the repo as a whole becomes an ungoverned patchwork of trust settings, hooks, MCP coverage, and command loops.

Every major AI coding platform

Works across the major AI coding surfaces teams already mix in one repo.

Claude Code
400 checks
Codex
272 checks
Gemini CLI
300 checks
Copilot
299 checks
Cursor
301 checks
Windsurf
297 checks
Aider
283 checks
OpenCode
286 checks

How it works

You do not need every product surface on day one. These three steps get you to the first useful answer fast.

01

Detect cross-platform drift

Find contradictory instructions, missing MCP coverage, permission gaps, and tool-specific files that have drifted out of alignment.

$ nerviq harmony-audit
02

Score the live repo

Run the current repo through the full governance audit so you can separate baseline hygiene from cross-platform alignment.

$ nerviq audit
03

Align with reviewable changes

Generate a safe plan, inspect the diff, and close drift without blind overwrite behavior. Rollback and proof stay in the workflow.

$ nerviq plan --dry-run

Go deeper after the first pass

Once the repo is no longer drifting, these are the next layers worth opening up.

Harmony

Go deeper on multi-tool drift after the first pass. Harmony is the stable surface for cross-platform alignment.

Governance

Add team rules, permissions, hooks, and reviewable JSON policy surfaces once the basics are in place.

API + SDK

Automate audits and Harmony in internal workflows when the CLI is no longer enough on its own.

Proof surfaces

Use case studies and the public before/after repo when you need inspectable proof instead of marketing-only claims.

See it in action

One command gets you a live briefing on the repo. The rest of the workflow stays reviewable.

nerviq audit

$ nerviq audit

nerviq v1.30.0 — auditing project...

Detected: Next.js + TypeScript + Claude + Cursor

Harmony Score: 78/100 — 3 drift issues across 2 platforms (Claude Code + Cursor)

Run nerviq harmony-audit for the full cross-platform report.

Platform: Claude Code

CLAUDE.md .............. 23/25 checks passed

.claude/settings.json .. 6/9 checks passed

Platform: Cursor

.cursor/rules/ ......... 14/16 checks passed

.cursorrules ........... 4/7 checks passed

Drift: trust-mode mismatch (Claude relaxed / Cursor strict)

MCP coverage gap — 3 servers in Claude, 1 in Cursor

format-on-save hook missing from Cursor

3 suggestions found. Run nerviq harmony-sync to align.

Completed in 1.4s_

Works with every stack

JavaScript, Python, Go, .NET, mobile, and more.

JavaScript / TypeScript
Python
Go
Rust
Java
Ruby
.NET / C#
PHP
Flutter / Dart
Swift / Kotlin

Open source. Production ready.

AGPL-3.0 licensed. 475 tests. Zero runtime dependencies. Built to stay around after the first audit, not just the first demo.

AGPL-3.0475 tests0 runtime depsJavaScript