tinyBigGAMES

tinyBigGAMES / VindexLLM

Public

VindexLLM is a pure Delphi, GPU-powered LLM inference engine that uses Vulkan compute shaders to run GGUF models entirely on the GPU. It performs full transformer inference without relying on Python, CUDA, or other external runtimes, requiring only vulkan-1.dll, which is typically included with modern GPU drivers.

delphi llama-cpp llm-inference object-pascal ollama
16
4
100% credibility
Found Apr 18, 2026 at 16 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Pascal
AI Summary

VindexLLM is a Windows program that runs AI language models from standard files directly on your graphics hardware for fast text generation without Python or special toolkits.

How It Works

1
🔍 Discover VindexLLM

You stumble upon VindexLLM, a clever way to chat with powerful AI right on your own Windows computer using its built-in graphics power, no extra complicated software needed.

2
📥 Grab an AI personality

You download one of the ready-tested AI model files from the safe links provided, like a digital brain ready to think and talk.

3
🛠️ Set up the chat tool

You open the program files in your Delphi app, tweak the path to your AI file, build it once, and everything is ready to go.

4
💬 Ask your first question

You type a fun prompt like 'Explain how a computer works' and hit start, feeling the excitement as the AI wakes up.

✨ See the magic unfold

Word by word, the AI streams back smart, helpful responses super fast, all powered by your computer's graphics, and you can keep chatting endlessly.

Sign up to see the full architecture

3 more

Sign Up Free

Star Growth

See how this repo grew from 16 to 16 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is VindexLLM?

VindexLLM is a pure Delphi engine for full GPU-powered LLM inference, running standard GGUF models entirely on the GPU via Vulkan compute shaders. It handles everything from token embedding to attention layers and sampling without Python, CUDA, or external runtimes—just vulkan-1.dll included with modern GPU drivers. Developers get a simple API to load a model, feed a prompt, and stream generated tokens back.

Why is it gaining traction?

It ditches CUDA's NVIDIA lock-in and massive installs for Vulkan's cross-GPU compatibility on NVIDIA, AMD, and Intel hardware. Self-contained binaries start instantly with memory-mapped GGUF loading and minimal PCIe transfers, delivering 24 tok/s generation on a 3060 without dependency hell. The hook is embedding production-ready LLM inference into Delphi apps with zero setup.

Who should use this?

Delphi developers building Windows desktop tools that need local LLM inference without bloating installs. Game devs at tinyBigGAMES-style studios integrating chat AI into Vulkan-rendered apps. Experimenters testing GGUF models on non-NVIDIA GPUs for edge deployment.

Verdict

Early alpha with 16 stars and 1.0% credibility score—docs are solid but limited to Gemma 3 4B models, so expect tweaks for broader use. Worth watching or forking if you're in Delphi/Vulkan; skip for production until more architectures land.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.