aallam/openai-kotlin
Openai Kotlin
This project is a multiplatform software development kit designed to integrate large language models into Kotlin applications. It provides an asynchronous network client and a suite of orchestration tools that enable developers to connect applications to intelligence services, manage conversation threads, and handle structured data exchange.
The library is built as a Kotlin Multiplatform package, allowing for consistent code sharing across Android, iOS, and desktop environments. It utilizes structured concurrency to manage non-blocking network requests and real-time streaming responses, ensuring that long-running operations do not interfere with application performance.
The toolkit covers the full lifecycle of AI-driven features, including the management of assistant configurations, conversation history, and external tool definitions. It supports function calling by allowing models to trigger external actions and submit results back to active execution runs. Security is maintained through integrated authentication mechanisms that manage credential injection for outgoing service requests.
Features
- OpenAI API Clients - Integrates large language models into applications by implementing the service API specification for sending prompts and receiving model responses.
- Assistant Thread Runners - The OpenAI client library processes messages within a conversation thread using a specific assistant and tracks the execution status until the model finishes its assigned work.
- Chat Completion Services - Generates conversational responses using structured message sequences with support for real-time streaming.
- Assistant Lifecycle Management - Provides operations for creating, retrieving, and updating AI assistant instances and their configurations.
- External API Tool Exposures - Exposes action functions as callable tools so models can fetch live data or perform external actions.
- AI Assistants - Manages the lifecycle of AI assistants and conversation threads to automate complex tasks.
- Function Calling Interfaces - Implements interfaces for language models to interact with live data through external tool calls.
- Conversation Threads - Maintains persistent containers for organizing user-agent message history across multiple interactions.
- Tool Output Submissions - Submits the results of tool calls back to an active assistant run to continue processing.
- Multiplatform Libraries - Provides a cross-platform software development kit that enables consistent code sharing across Android, iOS, and desktop environments.
- Multiplatform Serialization Libraries - Maps complex JSON responses into type-safe objects across multiple Kotlin targets.
- AI Assistant SDKs - Provides a toolkit for managing conversation threads, tool execution, and assistant lifecycles.
- Cross-Platform Application Development - Enables building intelligent features that run consistently across mobile and desktop environments using shared code.
- Asynchronous Network Clients - Provides a non-blocking communication interface for structured data exchange and streaming responses.
- Chunked Response Processing - Processes incoming server-sent event chunks to provide real-time updates.
- Authenticated Request Clients - Provides a modular client for authenticated HTTP communication and request lifecycle management.
- Coroutine-Based Asynchronous I/O - Uses structured concurrency to handle non-blocking network requests and streaming operations.
- Programming Libraries - Kotlin API client with multiplatform and coroutine support.
