tesvik-ai-cli

Autonomous AI-Powered Testing CLI

Automate your entire QA pipeline with one command — no Python required.

Tesvik.AI ingests any codebase — local folder or GitHub URL — analyzes its architecture, then autonomously generates, executes, and reports on a comprehensive test suite across 15 testing disciplines. Powered by Claude and Gemini.

Install

npm install -g tesvik-ai-cli

Requirements: Node.js 18+, and either ANTHROPIC_API_KEY or GOOGLE_API_KEY set in your environment.

Quick Start

export ANTHROPIC_API_KEY=sk-ant-...

tesvik-ai analyze ./my-project
tesvik-ai run ./my-project --type unit,api,security

Commands

tesvik-ai analyze <path>

Analyze a project and display its detected stack.

# Analyze a local project
tesvik-ai analyze ./my-project

# Analyze a GitHub repository
tesvik-ai analyze https://github.com/owner/repo

# Analyze with interactive AI Q&A + test plan
tesvik-ai analyze ./my-project --interactive

# Use Gemini instead of Claude
tesvik-ai analyze ./my-project --provider gemini

# Specific branch
tesvik-ai analyze https://github.com/owner/repo --branch develop

tesvik-ai run <path>

Generate and run tests for a project.

# Generate and run unit tests
tesvik-ai run ./my-project --type unit

# Run multiple agents
tesvik-ai run ./my-project --type unit,api,security

# Run all 15 agents
tesvik-ai run ./my-project --type all

# Generate tests only — skip execution
tesvik-ai run ./my-project --type all --no-run

# Custom output directory
tesvik-ai run ./my-project --type all --output my-tests

tesvik-ai ci <target>

Generate a GitHub Actions workflow.

tesvik-ai ci ./my-project
tesvik-ai ci ./my-project --type smoke,unit,api,security

tesvik-ai report

Print paths for the latest generated reports.

tesvik-ai report
tesvik-ai report --output my-tests

Available Agents (--type)

| Agent | Discipline | |---|---| | unit | Unit Testing | | api | API Testing | | integration | Integration Testing | | security | Security (SAST + DAST) | | e2e | End-to-End (Playwright) | | bulk_data | Bulk Data Testing | | performance | Performance Testing | | stress | Stress & Chaos | | scalability | Scalability Testing | | contract | Contract Testing | | fuzzing | API Fuzzing | | mutation | Mutation Testing | | accessibility | Accessibility Testing | | smoke | Smoke & Sanity | | regression | Regression Testing |

Use all to run every agent:

tesvik-ai run ./my-project --type all

API Keys

Set your AI provider key as an environment variable:

# Claude (Anthropic)
export ANTHROPIC_API_KEY=sk-ant-...

# Gemini (Google)
export GOOGLE_API_KEY=AI...

# NVIDIA — free tier, uses Qwen3-Coder (great for code generation)
export NVIDIA_API_KEY=nvapi-...

You only need one key to get started. The CLI automatically tries the next provider if the primary fails or hits a rate limit.

| Provider | Model | Get Key | |---|---|---| | claude | claude-sonnet-4-6 / opus-4-6 | console.anthropic.com | | gemini | gemini-2.5-pro | aistudio.google.com | | nvidia | qwen3-coder-480b (free) | build.nvidia.com |

Switch provider at runtime:

# CLI flag
tesvik-ai run ./my-project --type all --provider nvidia

# Interactive mode
/provider nvidia

Output

Generated tests are written to tesvik-tests/ (or your --output directory):

tesvik-tests/
├── unit/           # Generated unit test files
├── api/            # Generated API test files
├── security/       # Generated security test files
└── results/
    ├── report.html # Visual HTML report
    ├── report.json # Machine-readable report
    └── report.md   # Markdown summary

Architecture

src/
├── cli.ts                    # CLI entry point (analyze, run, ci, report)
├── core/
│   ├── models.ts             # CodeSnapshot, SourceFile interfaces
│   ├── localReader.ts        # Walk local directories
│   ├── githubReader.ts       # Clone & read GitHub repos
│   └── languageDetector.ts   # Detect languages & frameworks
├── ai/
│   └── router.ts             # Claude + Gemini router with retry/fallback
├── agents/                   # 15 test generation agents
├── execution/
│   ├── runner.ts             # pytest / jest subprocess runner
│   ├── dependencyInstaller.ts
│   └── resultParser.ts       # JUnit XML + Jest JSON parser
├── orchestrator/
│   └── qaSession.ts          # Interactive Q&A + test plan
└── reporting/
    └── generator.ts          # HTML + JSON + Markdown reports

Tech Stack

• Runtime: Node.js 18+ (TypeScript)
• AI Models: Anthropic Claude + Google Gemini
• CLI: Commander + Chalk + Ora + Boxen
• Git: simple-git (GitHub clone)
• Parsing: fast-xml-parser (JUnit XML)
• Test execution: execa (pytest / jest subprocess)

License

Proprietary — All rights reserved © 2026 Tesvik.AI

This software is proprietary and confidential. It is distributed exclusively through npm for end-use purposes. Copying, modifying, redistributing, or reverse engineering the source code is strictly prohibited without written permission from Tesvik.AI.

For licensing inquiries: [email protected]