coverit

Your code, covered. One command.

What is coverit?

coverit is an AI-powered test generation and execution platform. It analyzes your code changes, determines what tests are needed, generates them, runs them, and reports the results. One command: coverit run.

Quick Start

# Install globally
npm install -g @devness/coverit

# Or run with npx (no install)
npx @devness/coverit run

# Scan first (dry run)
coverit scan

# Full pipeline
coverit run

Features

• Smart diff analysis -- parses git diffs to identify exactly what changed: new endpoints, modified functions, added components.
• Automatic test strategy -- decides which test types to generate (unit, API, E2E, mobile, desktop) based on file types and framework detection.
• Multi-framework support -- generates tests for Vitest, Jest, Playwright, Detox, and more, using the correct idioms for each.
• Framework detection -- identifies Hono, Express, NestJS, React, Expo, Tauri, and other frameworks from your project config.
• Parallel execution -- organizes tests into execution phases and runs independent plans concurrently.
• Coverage tracking -- collects line, branch, function, and statement coverage from test runners.
• MCP integration -- use coverit directly inside Claude Code as a set of MCP tools.
• Local + Cloud -- execute tests locally or in cloud sandboxes (E2B, Docker, Hetzner).

How It Works

git diff → analyze → strategize → generate → execute → report

Step	What happens
Diff	Parses git changes to identify modified files, hunks, and line ranges.
Analyze	Scans source code to extract functions, classes, endpoints, components, and dependencies.
Strategize	Builds a TestStrategy with prioritized TestPlan items and phased execution order.
Generate	Produces test files using framework-specific generators (unit, API, E2E, mobile, desktop).
Execute	Runs generated tests via local runners, browsers, simulators, or cloud sandboxes.
Report	Aggregates results into a CoveritReport with pass/fail counts, coverage, and failure details.

CLI Reference

Command	Description
coverit scan [path]	Analyze changes and display the test strategy without generating or running tests.
coverit generate [path]	Generate test files based on the analysis, but do not execute them.
coverit run [path]	Full pipeline: analyze, generate, execute, and report.
coverit report	Display the results of the last run.

CLI Options

Option	Description
--type <type>	Restrict to a specific test type: unit, api, e2e-browser, e2e-mobile, e2e-desktop.
--env <env>	Execution environment: local, cloud-sandbox, browser, mobile-simulator, desktop-app.
--coverage	Collect coverage data during execution.
--dry-run	Show what would happen without writing files or running tests.
--verbose	Enable debug-level logging (COVERIT_DEBUG=1).

MCP Integration

Add coverit as an MCP server in your Claude Code config:

{
  "mcpServers": {
    "coverit": {
      "command": "npx",
      "args": ["@devness/coverit", "mcp"]
    }
  }
}

MCP Tools

Tool	Description
coverit_analyze	Analyze code changes and return the test strategy.
coverit_generate	Generate test files from a strategy.
coverit_run	Execute generated tests and return results.
coverit_full	Full pipeline: analyze, generate, execute, report.

Supported Frameworks

Category	Frameworks
API	Hono, Express, NestJS, Fastify
Frontend	React, Next.js
Mobile	React Native, Expo
Desktop	Tauri, Electron
Test Runners	Vitest, Jest, Playwright, Detox, Pytest
Package Managers	Bun, pnpm, npm, yarn

Architecture

┌──────────────┐
│  CLI / MCP   │  ← Entry points
└──────┬───────┘
       │
       ▼
┌──────────────┐
│ Orchestrator  │  ← Coordinates pipeline
└──────┬───────┘
       │
       ├──────────────┬──────────────┐
       ▼              ▼              ▼
┌────────────┐ ┌───────────┐ ┌────────────┐
│  Analysis   │ │ Generators │ │  Executors  │
│  Engine     │ │            │ │             │
│ - Diff      │ │ - Unit     │ │ - Local     │
│ - Scanner   │ │ - API      │ │ - Cloud     │
│ - Deps      │ │ - E2E      │ │ - Browser   │
│ - Strategy  │ │ - Mobile   │ │ - Simulator │
└────────────┘ │ - Desktop  │ └──────┬──────┘
               └───────────┘        │
                                    ▼
                             ┌────────────┐
                             │  Reporter   │
                             └────────────┘

See docs/architecture.md for a detailed breakdown.

Configuration

Create a coverit.config.ts in your project root:

import { defineConfig } from "coverit";

export default defineConfig({
  projectRoot: ".",
  testTypes: ["unit", "api", "e2e-browser"],
  environment: "local",
  coverageThreshold: 80,
});

See docs/configuration.md for all options.

Roadmap

• Cloud execution (E2B, Docker)
• AI-powered test refinement on failure
• PR comment integration (GitHub Actions)
• VS Code extension
• Coverage trend tracking across runs

Contributing

See CONTRIBUTING.md for development setup and guidelines.

License

MIT