AI Image Generation from Delphi — A 2026 Roundup

The Delphi ecosystem has grown significantly for AI integration. Here’s what’s available today for adding AI image generation to your Delphi applications:

Official Components (GetIt / RAD Studio)

Stability.ai API Wrapper for Delphi — Available on GetIt. Wraps Stability AI’s cloud API for Stable Diffusion 3.5, Stable Image Ultra. Supports image generation, editing, upscaling, and outpainting. The most straightforward path for cloud-based image generation in Delphi.

OpenAI for Delphi — Available on GetIt. Full OpenAI API access including GPT Image 1.5 (successor to DALL-E 3, released December 2025). Covers text, vision, audio, and image endpoints.

SmartCore AI Components Pack — Included in RAD Studio 13+. Multi-provider (OpenAI, Claude, Gemini, Ollama). Features TAIImageRequest component with visual designer support. Good starting point for apps that need to support multiple AI providers.

MakerAI Suite — Available on GetIt and GitHub. Enterprise-grade, 100% native Delphi. Supports OpenAI, Claude, Gemini, Groq, Ollama, Mistral. Includes RAG, autonomous agents, and vision capabilities. Most feature-complete option.

Community Libraries (Open Source)

MaxiDonkey’s collection is the standout. This developer maintains high-quality wrappers for nearly every major AI provider:

  • DelphiStabilityAI — Stability AI (image gen, editing, outpainting)
  • DelphiGenAI — OpenAI (GPT Image 1.5, SORA-2 video)
  • DelphiHuggingFace — Access to thousands of models including SD variants
  • DelphiGemini — Google Gemini with vision capabilities
  • DelphiAnthropic — Claude (analysis, not generation)

All on GitHub, open source, actively maintained.

HemulGM/DelphiOpenAI — Mature, widely-used OpenAI wrapper. Also supports DeepSeek, Azure OpenAI, YandexGPT, Ollama, and more. No external dependencies.

FMXExpress/Stable-Diffusion-Desktop-Client — Complete reference implementation of a Delphi app that generates images via Replicate’s API. Cross-platform (FMX). Great learning resource.

Commercial Components

TMS AI Studio (TMS Software) — TTMSFNCCloudImageAI component. Part of the TMS All-Access subscription. Connects to ChatGPT, Claude, Gemini. Integrates with TMS FNC UI Pack. Expensive? Too big? TMS libraries are usually big.

For Local AI (ComfyUI, SD WebUI / Forge)

No Delphi-specific libraries exist for local SD/ComfyUI integration. However, both expose standard HTTP/REST APIs:

  • ComfyUI runs on port 8188 with a queue-based API (POST workflow JSON, poll for results). Actively maintained with major 2026 updates: 2.5x faster on RTX 50 series, 60% lower VRAM usage, new “App View” for beginners.
  • SD WebUI / Forge runs on port 7860 with synchronous endpoints (/sdapi/v1/img2img for inpainting). Note: Forge has largely superseded the original Automatic1111 with 6x better performance and identical API. SD.Next is another compatible fork.

Delphi’s built-in THTTPClient, TMultipartFormData, and TJSONObject are all you need. No external library required.

The Bottom Line

For cloud AI image generation, Delphi is well-served. The Stability.ai wrapper on GetIt or MaxiDonkey’s open-source libraries get you running in minutes.

For local AI (ComfyUI, Forge), you’re writing HTTP calls yourself — but it’s straightforward REST, and the built-in Delphi HTTP stack handles it well.

The gap is in high-level abstractions — there’s no single Delphi component that unifies local and cloud providers behind one interface. That’s something each application needs to build. BioniX is building exactly this: a provider abstraction layer that lets users choose between local Stable Diffusion and cloud APIs like Stability AI, all through the same settings panel.

 

Leave a Comment

Scroll to Top