Developer MaxiDonkey has released DelphiGenAI, a native OpenAI integration framework for Delphi that brings modern multimodal AI, realtime interactions, agent workflows, and next-generation OpenAI APIs directly into VCL and FireMonkey applications.
The library provides support for OpenAI Responses APIs, GPT-5 and reasoning models, embeddings, vector stores, vision workflows, speech processing, realtime streaming, image generation, and Sora video generation through a strongly typed Object Pascal API.
The framework includes support for:
- Agentic AI workflows and Responses APIs
- GPT-5, o3, and multimodal models
- Vision, audio, and speech processing
- Realtime AI streaming
- Image and Sora video generation
- Embeddings and vector stores
- Structured outputs and function calling
- Async and parallel AI workflows
DelphiGenAI also includes both VCL and FireMonkey demo applications, making it easy to test modern AI workflows across desktop and cross-platform Delphi projects.
Built for Windows, macOS, Linux, Android, and iOS, the framework supports OpenAI-compatible providers including OpenAI, Gemini, Claude, DeepSeek, Grok, and local LM Studio deployments.
Explore DelphiGenAI, the VCL and FMX demos, and the full framework.