Skip to main content
Esprit Hero Light

​
Welcome to Esprit

Esprit is an autonomous security assessment tool that uses AI agents to perform comprehensive penetration tests. Unlike traditional scanners, Esprit thinks like a human pentester—exploring your application, analyzing behavior, and discovering complex vulnerabilities that automated tools miss.

Installation

Install Esprit using curl, Homebrew, npm, pip, or from source

Quick Start

Run your first security scan in under 5 minutes

Scan Modes

Choose between quick, standard, and deep scanning strategies

Provider Setup

Connect LLM providers for cloud or local scanning

​
How It Works

Esprit combines autonomous AI agents with security testing tools in an isolated sandbox environment: The agent observes your application’s behavior, crafts targeted attacks, analyzes responses, and iteratively explores until it maps out your security posture.

​
Key Features

Test web applications, APIs, GitHub repositories, and local codebases. Esprit automatically detects target types and adapts its testing strategy.
esprit scan https://api.example.com
esprit scan https://github.com/user/repo
esprit scan ./my-project
Esprit discovers:
  • SQL Injection & NoSQL Injection
  • Cross-Site Scripting (XSS)
  • Authentication & JWT Flaws
  • IDOR & Broken Access Control
  • SSRF & Path Traversal
  • Race Conditions
  • Business Logic Vulnerabilities
  • Mass Assignment & CSRF
  • Open Redirects & Information Disclosure
Esprit Cloud: Run scans without Docker using your subscriptionLocal Mode: Use any LLM provider with Docker for the sandboxBoth modes deliver the same comprehensive security testing.
Vulnerabilities are displayed as they’re discovered, with:
  • CVSS severity scores
  • Proof-of-concept exploits
  • Remediation guidance
  • HTTP request/response evidence
All results are saved to esprit_runs/<scan-id>/ for later review.

​
Vulnerability Detection

Esprit tests for a comprehensive range of security issues:

Injection Flaws

SQL, NoSQL, OS command, LDAP, XPath injection

Authentication

Broken auth, session management, JWT flaws

Access Control

IDOR, privilege escalation, broken authorization

XSS & CSRF

Reflected, stored, DOM-based XSS, CSRF tokens

Server-Side

SSRF, XXE, file inclusion, path traversal

Business Logic

Race conditions, workflow bypasses, mass assignment
Authorized Testing OnlyEsprit is designed for authorized security testing. Only test systems you own or have explicit written permission to test.

​
Scan Modes

Choose a scanning strategy based on your time and coverage needs:
ModeDurationUse Case
Quick~5 minFast surface-level scan for CI/CD pipelines
Standard~15 minBalanced coverage for regular testing
Deep~30+ minComprehensive analysis for thorough audits
esprit scan https://example.com -m quick      # Fast
esprit scan https://example.com -m standard   # Balanced
esprit scan https://example.com -m deep       # Thorough (default)

​
Architecture

Esprit runs your target application inside an isolated Docker container (or cloud sandbox) equipped with:
  • Caido proxy for intercepting and analyzing HTTP traffic
  • Playwright for browser automation and DOM manipulation
  • Security tools like nmap, sqlmap, nuclei, ffuf
  • Python runtime for custom exploitation scripts
  • File editor for analyzing and modifying code
The sandbox prevents malicious code from escaping, but always review scan results before taking action in production.

​
Next Steps

1

Install Esprit

Choose your preferred installation method: curl, Homebrew, npm, pip, or source
2

Connect a Provider

Set up Esprit Cloud for zero-config scanning, or configure your own LLM provider
3

Run Your First Scan

Follow the Quick Start guide to scan your first target
4

Review Results

Analyze vulnerability reports and implement fixes

Need Help?

Report issues or ask questions on GitHub

Build docs developers (and LLMs) love

Get started for freeTalk to us