AgenticKode Documentation

Welcome to the AgenticKode documentation.

Getting Started

• Installation
• Quick Start
• Configuration

How runs work

A run is agentic: you give an agent a prompt and it runs to completion — analyzing the repo, writing code, running tests, and opening a PR. AgenticKode auto-fetches the context (repo, issue body, PR diff) so the agent starts grounded.

• Autonomous Mode — give an agent a task; it runs end-to-end and opens a PR
• Controlling with AI Agents — conversational chat + agent-to-agent collaboration via MCP

Guides

• Webhook Setup — fire runs from GitHub, GitLab, Gitea, Plane, Notion events
• Workspace Servers