Can I use Koog in Spring Boot applications?

Yes, Koog is compatible with Spring Boot and provides a specific Ktor integration module (koog-ktor) for seamless backend implementation.

Which LLM providers does Koog support?

Koog supports a wide range of providers including OpenRouter, OpenAI, Anthropic, Google Gemini, DeepSeek, and local Ollama instances.

How does Koog handle malformed JSON from LLMs?

Koog includes a StructureFixingParser that attempts to parse responses into typed data classes and can automatically trigger a secondary LLM call to fix formatting errors if they occur.

What is the Koog AI Agent framework?

Koog is a specialized Kotlin framework designed to simplify the creation of AI agents that can call tools, reason through complex tasks, and integrate with multiple LLM providers.

Koog AI Agent Framework

Name: Koog AI Agent Framework
Author: AndVl1

byAndVl1

•

数据科学与机器学习

Builds sophisticated AI agents with tool-calling capabilities and multi-provider LLM integration using a Kotlin-native framework.

Koog is a powerful Kotlin Multiplatform framework designed for developing advanced AI agents and LLM-powered workflows within the JetBrains ecosystem. It offers a comprehensive suite of tools including a type-safe Prompt DSL, annotation-based tool definitions, and flexible reasoning strategies like ReAct and Chat loops. By providing native support for major providers such as OpenRouter, Anthropic, and OpenAI—alongside specialized handling for structured JSON output—Koog enables developers to build, test, and deploy production-grade AI features with type safety and architectural clarity.

主要功能

01Annotation-based tool registry for seamless LLM function calling

02Flexible strategy DSL for custom reasoning loops and state machines

03Kotlin-native agent construction with multiplatform support

04Multi-provider support including OpenRouter, Anthropic, and DeepSeek

051 GitHub stars

06Structured output parsing with automated LLM-based error correction

使用场景

01Building autonomous AI agents that interact with local file systems or internal APIs

02Creating complex multi-step reasoning workflows using custom Graph strategies

03Integrating type-safe structured LLM responses into Spring Boot or Ktor backends

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add andvl1/claude-plugin koog

For use in Claude.ai and ChatGPT

Download Skill