Skip to main content

​
Introduction

PentAGI provides a comprehensive GraphQL API for managing autonomous penetration testing workflows, AI agents, and security assessments. The API enables you to create and manage flows, tasks, assistants, and access detailed execution logs.

​
Endpoint

The GraphQL API is available at: 
POST /api/v1/graphql

​
GraphQL Playground

An interactive GraphQL playground is available for testing queries: 
GET /api/v1/graphql/playground

​
WebSocket Endpoint

For real-time subscriptions, connect to the same endpoint using WebSocket protocol: 
ws://your-server/api/v1/graphql

​
Key Features

Flow Management

Create and manage penetration testing flows with AI-powered automation

Real-time Updates

Subscribe to live updates via GraphQL subscriptions

Agent Orchestration

Control specialized AI agents for different security tasks

Comprehensive Logging

Access detailed execution logs, terminal output, and agent communications

​
Core Concepts

​
Flows

Flows represent complete penetration testing sessions. Each flow:
  • Has a unique ID and title
  • Executes with a specific LLM provider
  • Contains multiple tasks and subtasks
  • Tracks execution status (created, running, waiting, finished, failed)
  • Manages isolated Docker containers (terminals)

​
Tasks & Subtasks

Tasks are high-level objectives within a flow. Each task:
  • Can be broken down into subtasks
  • Has input, result, and status tracking
  • Logs all agent interactions and tool calls

​
Assistants

Assistants are interactive AI agents that can:
  • Answer questions about the flow
  • Execute follow-up tasks
  • Optionally use the full agent team or work independently

​
Providers

PentAGI supports multiple LLM providers:
  • OpenAI
  • Anthropic
  • Google Gemini
  • AWS Bedrock
  • Ollama (local)
  • Custom endpoints

​
Response Format

All GraphQL responses follow the standard format: 
{
  "data": {
    // Requested data
  },
  "errors": [
    // Optional errors array
  ]
}

​
Rate Limits

Rate limits depend on your deployment configuration and chosen LLM provider limits.

​
Next Steps

Authentication

Learn how to authenticate API requests

Flows API

Manage penetration testing flows

Tasks API

Work with tasks and subtasks

Agents API

Control AI agents and view logs

Build docs developers (and LLMs) love

Get started for freeTalk to us