Open Issues Need Help
View All on GitHubKoog is the official Kotlin framework for building predictable, fault-tolerant and enterprise-ready AI agents across all platforms – from backend services to Android and iOS, JVM, and even in-browser environments. Koog is based on our AI products expertise and provides proven solutions for complex LLM and AI problems
AI Summary: The user needs help understanding how to use the RAG (Retrieval Augmented Generation) components within the Koog framework, specifically the `vector-storage` module. The provided `build.gradle.kts` snippet shows dependencies for other Koog modules, but the RAG dependencies and how to include them in a project using version 0.2.1 are unclear. The user needs clarification on the correct dependencies and how to integrate the `vector-storage` module for RAG functionality.
Koog is the official Kotlin framework for building predictable, fault-tolerant and enterprise-ready AI agents across all platforms – from backend services to Android and iOS, JVM, and even in-browser environments. Koog is based on our AI products expertise and provides proven solutions for complex LLM and AI problems
AI Summary: The issue describes an incompatibility between the specified OpenAI model (OpenAIModels.CostOptimized.GPT4_1Mini) and the available models supported by the Grazie LLM client used within the Koog framework. The task is to debug this mismatch, potentially by updating the Koog framework's model compatibility list or adjusting the agent's configuration to use a supported model. This might involve investigating the Grazie LLM client's API documentation and potentially contributing a fix or improvement to the Koog framework.
Koog is the official Kotlin framework for building predictable, fault-tolerant and enterprise-ready AI agents across all platforms – from backend services to Android and iOS, JVM, and even in-browser environments. Koog is based on our AI products expertise and provides proven solutions for complex LLM and AI problems
AI Summary: The task is to investigate and potentially refactor the tool annotation parsing in the Koog framework. Currently, it uses reflection and ignores a provided JSON parameter, leading to inconsistencies and limitations in handling certain data types. The solution involves leveraging kotlinx.serialization for improved type handling and consistency between serialized values and descriptions.
Koog is the official Kotlin framework for building predictable, fault-tolerant and enterprise-ready AI agents across all platforms – from backend services to Android and iOS, JVM, and even in-browser environments. Koog is based on our AI products expertise and provides proven solutions for complex LLM and AI problems
AI Summary: Rename the `TextEmbedding3Small` and `TextEmbedding3Large` constants in the `OpenAIModels.kt` file to reflect a more consistent and descriptive naming convention. This likely involves a simple find-and-replace operation, potentially with minor adjustments to ensure code clarity.
Koog is the official Kotlin framework for building predictable, fault-tolerant and enterprise-ready AI agents across all platforms – from backend services to Android and iOS, JVM, and even in-browser environments. Koog is based on our AI products expertise and provides proven solutions for complex LLM and AI problems
AI Summary: The task is to add support for specifying embedding dimensions in the Koog framework's OpenAI and Ollama integrations. This involves adding a `dimensions` property to the relevant embedding request objects and verifying Ollama's support for this feature.
Koog is the official Kotlin framework for building predictable, fault-tolerant and enterprise-ready AI agents across all platforms – from backend services to Android and iOS, JVM, and even in-browser environments. Koog is based on our AI products expertise and provides proven solutions for complex LLM and AI problems