Free AI tools that run on an 8 GB GPU

A web UI for generating and editing images using Stable Diffusion.

The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.

Generate images from text prompts with minimal effort.

Generate and refine images using Stable Diffusion models.

Generates images, videos, and audio using Diffusion models.

AI image generation and inpainting for Krita.

Speech recognition model for multilingual transcription and translation.

Clone a voice in real-time for generating speech.

Generates high-quality multilingual speech from text.

Transcribes audio faster than OpenAI Whisper with CTranslate2.

Convert e-books to audiobooks with text-to-speech and voice cloning.

Generate text-to-speech with emotional control and multiple voices.

Access and use Alibaba Cloud's Qwen large language models for chat and text generation.

RWKV-7 is a linear-time RNN language model for large-scale inference and chat applications.

An open-source conversational language model supporting Chinese and English.

Open-source chatbot powered by RWKV RNN language model.

AI research assistant for deep, agentic research using local LLMs and search engines.

Train a 65M-parameter visual-language model in 2 hours.

Reproduces DeepSeek R1-Zero for reinforcement learning in language models.

Trains and runs hierarchical reasoning models for complex tasks.

Train large language models from scratch using PyTorch.

Fine-tune LLMs with one command and minimal resources.

Toolkit for merging pre-trained language models.

Cross-vendor 3D Gaussian Splatting trainer - video to splat to mesh, Vulkan or CUDA.

Runs large language models on low VRAM GPUs.

A fast, lightweight inference engine for language models.

Self-host any open-source LLM as an OpenAI-compatible API server.

Run local AI apps and models on your GPU and NPU.

Rust-based OpenAI-compatible inference server for GGUF models.

Serve Qwen models on consumer GPUs with vLLM optimizations.

Accelerated video generation framework for real-time inference.

Generates human animations from audio inputs.

Generates high-fidelity video from text prompts.

Generates detailed visuals for playable games in real-time.

Generates high-quality human motion videos with confidence-aware pose guidance.

Generate videos using local hardware with MiniMax H3.

Real-time neural graphics primitives for NeRF, SDF, and image rendering.

提取视频中的硬字幕并生成srt文件。

ML-powered manga translator with OCR and inpainting for local manga translation.

Trains and tests monocular depth estimation models using PyTorch.

Rust OCR engine with multiple backends and an OpenAI-compatible server.

An local, offline (after initial setup), portable OCR software that can process images and PDF files, using DeepSeek-OCR-2 AI (running directly on your machine).

Runs large language models on low-memory Macs using Swift and Metal.

Portable AI music generator — full songs with vocals, covers, music videos. One-click install, 100% offline, NVIDIA GPU.

Hybrid RAG system combining vector search, knowledge graph (LightRAG), and cross-encoder reranking — with Docling document parsing, visual intelligence (image/table captioning), agentic streaming chat

An AI-powered natural language & reverse Image Search Engine powered by CLIP & qdrant.

The black box for AI-written code. Passive decision graph linking every commit to the agent prompts behind it: drift detection, DLP audit of what agents read, evidence-driven risk, MCP memory for codi