Inkling⬇16kTrending Model:#2
Ternary-Bonsai-27B-gguf⬇432kTrending Model:#3
Unlimited-OCR⬇2237kTrending Model:#4Bonsai-27B-gguf⬇1405kTrending Model:#5
GLM-5.2⬇545kTrending Model:#6
Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-NEO-MAX-MTP-GGUF⬇63kTrending Model:#7
Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive⬇1998kTrending Model:#8
Qwythos-9B-Claude-Mythos-5-1M-GGUF⬇2133kTrending Model:#9
krea2-identity-edit⬇0kTrending Model:#10
OvisOCR2⬇17kTrending Model:#1Inkling⬇16kTrending Model:#2Ternary-Bonsai-27B-gguf⬇432kTrending Model:#3Unlimited-OCR⬇2237kTrending Model:#4Bonsai-27B-gguf⬇1405kTrending Model:#5GLM-5.2⬇545kTrending Model:#6Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-NEO-MAX-MTP-GGUF⬇63kTrending Model:#7Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive⬇1998kTrending Model:#8Qwythos-9B-Claude-Mythos-5-1M-GGUF⬇2133kTrending Model:#9krea2-identity-edit⬇0kTrending Model:#10OvisOCR2⬇17k
DeepSeek-v4-Fable is an autonomous agent engineered for offensive security research. This release operates as a specialized tool for solving challenges and planning exploits in controlled environments. It acts as a […]
Baidu has introduced Unlimited-OCR, a new tool designed to read and transcribe long documents without losing speed. This model processes dozens of pages in a single pass by maintaining a […]
ComfyUI-Yedp-UV-Painter is an experimental, non-destructive 3D-to-2D AI texturing node for ComfyUI. It bridges professional 3D pipelines with AI image generation directly inside the browser interface. Users can surgically texture complex […]
Supra-A2A-Nano-Exp is an experimental proof-of-concept any-to-any model that processes text, images, and video using a single system. It translates visual inputs into a small set of learned codes and treats […]
MiMo-Audio-7B-Instruct is a new audio language model designed to understand and generate sound based on simple instructions. It learns from a massive amount of audio data to perform tasks like […]
The new release called MiMo-Audio-7B-Base is an open-source audio language model designed to learn new tasks from just a few examples. It processes over one hundred million hours of audio […]
Boogu-Image-0.1-Edit is an open-source image editing and transformation model that works alongside a broader multimodal generation family. This release allows users to perform complex editing tasks like object insertion, background […]
The new release of Ideogram_4_turbotime_lora transforms the standard Ideogram 4 model into a fast image generator. It allows the system to create good quality images in as little as two […]
The new release of Lift provides a way to pull organized data out of PDFs and images. Users can provide a standard JSON format, and the model will generate matching […]
Qwable-3.6-27b is a new full checkpoint model fine-tuned from the unsloth/Qwen3.6-27B base. This release pushes the base model to produce more deliberate and structured responses for coding and technical reasoning […]
Krea 2 Turbo is a new text-to-image diffusion model that turns natural language descriptions into pictures. It uses a diffusion transformer architecture with twelve billion parameters to create detailed visuals […]
Krea 2 Raw is a text-to-image diffusion model that generates pictures from written descriptions. This release serves as a base checkpoint for developers who want to customize the model. Users […]
Inflect-Nano-v1 is a tiny English text-to-speech model that turns written words into spoken audio. It includes its own audio generator and uses less than five million parameters to function. The […]
GLM-5.2-GGUF is a newly released AI model format built for handling long and complex tasks. It provides a solid one million token context window so you can process very large […]
The new release Gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF provides a local coding and tool-using agent that runs entirely offline. It allows users to run a private artificial intelligence model that can read, reason, and […]
Project Sulfur is a new open source desktop application that lets you run AI coding agents entirely on your local machine. It allows the AI to read, write, edit, and […]
PonyExl3 is a new tool that lets you run advanced compressed language models on Apple Silicon computers. It translates a specific high quality model format called Exl3 so it works […]
Tower-Plus-72B-ultra-uncensored-heretic is a new artificial intelligence model that removes content restrictions from a system originally built for translation and multilingual tasks. This release takes an existing language model and strips […]
Nex-N2-mini-Turbo-Phase-Twin is a new release that brings a large language model to Intel Arc graphics cards. It packages the Qwen3.5-35B-A3B model into a single file that holds two different precision […]
The new Command-a-plus-05-2026-GGUF release provides multiple compressed versions of the command-a-plus-05-2026 model built by CohereLabs. These files allow users to run large AI text generation models locally on their own […]
Glimmer-1-Base is a newly released experimental language model designed to explore the absolute minimum scale needed for useful artificial intelligence. This tiny transformer features only 11,900 parameters and was trained […]
The new release of Trellis2-mlx brings a powerful image-to-3D generation pipeline to Apple Silicon computers. It allows users to take a standard 2D picture and turn it into a detailed […]
ComfyUI-MAMMA is a new set of custom nodes that brings markerless multi-camera motion capture into the ComfyUI interface. It allows users to process synchronized video to fit 3D body models […]
Zonos2_TTS-ComfyUI is a new custom node integration that brings advanced text-to-speech and audio-only voice cloning capabilities directly into ComfyUI workflows. The release allows users to generate high-fidelity speech from text […]
Turbo-LLM is a new tool that lets you run local language models with an automatic tuning setup for your graphics card. It provides a polished web interface and works with […]
ComfyUI-NKD-Klein-Tools simplifies complex Flux Klein workflows into a straightforward set of nodes. Users can plug in a model, drop an image, and write a prompt to generate or edit pictures […]
UltraReal_FineTune_Anima_base1_v3 is an updated AI image model designed to push realism further from its original anime base. This release improves overall aesthetic quality and generates a wider variety of faces, […]
ExpressionControl is a new collection of models that allows for fine-grained and linear control over facial expressions. It enables users to adjust specific emotions while maintaining high character consistency. This […]
The new neodragon release is a text-to-video generation system capable of creating two-second videos at high resolution directly on mobile hardware. It takes text prompts and transforms them into 49 […]
Boogu-Image is a new open-source collection of artificial intelligence models designed for generating and editing pictures. This release offers stable tools for high-quality text-to-image creation, fast generation, and bilingual text […]
The new release of ideogram_4_unconditional_lora provides a lightweight method to run the Ideogram 4 model on computers with limited memory. It works by replacing a massive nine billion parameter component […]
Qwable-v1 is an open-weights artificial intelligence model designed for agentic coding tasks. It processes instructions by thinking in explicit reasoning chains before taking actions. The system then outputs specific XML […]
FastContext-1.0-4B-SFT is a lightweight repository exploration subagent designed to assist large language model coding agents. It works by separating the task of exploring a codebase from the actual problem solving […]
VibeThinker-3B is a newly released open source artificial intelligence model designed to solve complex reasoning problems. It focuses on mathematics, coding, and science tasks where the final answer can be […]
GLM-5.2 is a newly released artificial intelligence model built to handle complex, long-term tasks. It provides a stable one million token context window so it can process massive amounts of […]
MiMo-Code is a terminal-native AI coding assistant that reads and writes code while managing Git commands. It uses a persistent memory system to maintain a deep understanding of projects across […]
The new release called Elato Music is a local-first macOS application that connects to an ESP32-S3 device to stream realtime generative music throughout your house. It uses an open music […]
The newly released Refiner engine helps developers transform raw robotics and multimodal information into clean datasets for machine learning training. This tool provides a single pipeline model to handle multimodal […]
The new tool called xdna-top is a terminal system monitor built specifically for AMD Ryzen AI processors. It provides a unified, real-time view of both the neural processing unit and […]
The new release called Gemma-4-31B-it-qat-q4_0-unquantized-uncensored-heretic is a modified version of the Gemma 4 language model designed to bypass safety filters. It uses quantization-aware training to maintain performance while reducing memory […]
PP-OCRv6 is a lightweight optical character recognition system that extracts text from images and PDFs. The new release updates the toolkit to convert documents into structured data for AI applications. […]
browser-use-wasm is a client-side tool that lets vision-language models automate web browsers without a backend server. It captures screenshots of a webpage and uses an AI model to decide what […]
Supra-Title-350M-exp-GGUF is a new experimental artificial intelligence model designed specifically to generate titles for chat conversations. Instead of using massive systems that require heavy computing power, this compact tool focuses […]
Contextspy is a new profiling tool that helps developers understand and optimize how artificial intelligence coding agents use context windows. It intercepts requests sent to language model APIs to analyze […]
Open-dungeon is a fully local AI roleplay application that generates both text stories and inline scene images directly on your computer. It works entirely offline without needing any accounts, API […]
Fable-5-traces is a newly released dataset that captures the behavior of the Fable 5 coding agent. This collection converts original agent logs into a format that allows users to inspect […]
VISTA-9B is a visual model that understands screen layouts and translates natural language instructions into precise click coordinates. It looks at a screenshot and figures out exactly where to click […]
ZONOS2 is a new text-to-speech model designed to generate highly expressive and natural sounding audio. It predicts high quality audio tokens to create studio-grade sound at a 44.1 kHz sample […]
Kimi-K2.7-Code-GGUF is a coding-focused AI model designed to handle complex software engineering tasks from start to finish. It is built upon a previous version called Kimi K2.6 and improves token […]
The new pi-setup release provides a customized configuration for the Pi coding agent. It wraps an existing Pi installation to add extensions, custom themes, and sync tooling. Users can back […]
A new update to llama-launcher brings hands-free performance tuning to users of the llama.cpp local inference engine. The tool is a lightweight graphical application that lets you build server launch […]
SupraLabs has published Supra-1.5-50M-Base-exp, a continued pretraining update for their 50-million-parameter language model that stretches the usable context window fivefold. The release takes the original Supra-50M architecture and uses RoPE […]
Orion4D_FXMax is a new post-production suite for ComfyUI that packs color grading, sharpening, and film-like effects into a single set of nodes. It comes with a custom built-in web application […]
ComfyTV is a new canvas-based app that integrates natively with ComfyUI, turning the node-based interface into a visual workflow canvas. It lets you chain generative AI tasks like text-to-image, inpainting, […]
ComfyUI_PaletteDirector is a set of custom nodes that pairs visual color selection with a trained LoRA to steer image generation in ComfyUI. It builds a palette-card image and adds a […]
ComfyUI-Mutantwork is a new node pack that equips ComfyUI with prompt optimization, local AI forensic detection, and cryptographically signed content credentials. It strips out overused tokens that weaken results, analyzes […]
The Comfyui-ResolutionAndAspectRatio custom node for ComfyUI converts a chosen megapixel budget and aspect ratio into precise pixel dimensions. It can automatically detect the aspect ratio from words in your prompt, […]
Comfyui-PromptHistory is a new ComfyUI extension that adds two convenient nodes for saving every prompt you use during image generation. It stores those prompts in plain JSON files, making them […]
ComfyUI-Pin-Node-Input is a new ComfyUI extension that lets you pin any widget from any node into a sidebar panel. You can adjust prompt text, seeds, CFG, or sampler settings without […]
The SCAIL Auto Extend custom node for ComfyUI takes the manual labor out of generating long SCAIL-2 videos by handling every extension pass inside a single queue. It automatically breaks […]
A new plugin called Anomalous_Model_Browser has been released to solve a persistent headache for ComfyUI users: broken workflows when sharing projects. This is a lightweight, zero-dependency browser staying under 250KB […]
ComfyUI-PanTextHider is a lightweight extension for the ComfyUI node editor that hides text on nodes while you pan or zoom the canvas. It automatically fades out titles, labels, and widget […]
The LTX2.3-audio-reactive-lora is a LoRA adapter for the LTX-2.3 model that makes generated video move in sync with audio. The update released on June 15, 2026, sharpens how strongly visual […]
Comfyui-Focus-Mode is a new extension for ComfyUI that lets you isolate a handful of nodes into a clean, full-screen panel. It hides all other nodes and wires, giving you a […]
ComfyUI-JumpToNode is a new extension for ComfyUI that lets users instantly jump to any node by its ID, subgraph path, or keyword search. It helps navigate large, complex workflows where […]
Photoroom has released Prxpixel-t2i, a 7-billion-parameter text-to-image model that generates images directly in pixel space. Unlike most diffusion models, it does not use a variational autoencoder (VAE) to compress pixels […]
Sageattention-autotune is a fresh open source release that adds automatic block size tuning to the SageAttention 2 kernel. It brings CUDA and Triton kernel ports with quality-of-life improvements to speed […]
Dyfuzor-web is a tiny, self-contained web application that converts Excalidraw drawings into structured JSON prompts for the Ideogram AI image service. You draw a layout using familiar shapes and text, […]
SwiftVR is a new open-source AI model that cleans up and upscales low-quality video in real time, using a one-step generative approach to refresh each frame. It processes video in […]
The Comfyui-flow-upscaler node set brings a single-step latent upscaling method to ComfyUI, using a compact rectified flow model to enlarge Flux.2 images. It can upscale a 512x512 image to 1024x1024 […]
Ideogram-Json-Captioner is a new local desktop application that lets users edit and create structured JSON captions for image datasets. It provides a visual interface for manually editing Ideogram 4 fields […]
ComfyUI-Gradual-IC-LoRA is a new ComfyUI node pack that lets you gradually change the strength of an IC-LoRA effect over the course of a generated video. Instead of keeping a LoRA’s […]
ComfyUI-mnemic-nodes is a new pack of over 30 custom nodes for ComfyUI covering image generation, text tools, and metadata. It includes wildcard handling, LoRA loading, prompt parsing, and Groq API […]
TTS-Audio-Suite 5.3.0 has been updated to introduce a major architectural shift and expanded engine support. The open-source ComfyUI extension now runs on a Transformers 5 main environment while fragile engines […]
The Gemma-4-12B-coder-fable5-composer2.5-v1-GGUF is a quantized coding model that fits on low-resource hardware. It runs entirely offline without any cloud API, needing only around 4.5 GB of VRAM or unified memory. […]
Unsloth has released gemma-4-26B-A4B-it-qat-GGUF, a new quantized version of Google DeepMind’s Gemma 4 26B Mixture-of-Experts model. It uses Quantization-Aware Training to preserve near-original quality while shrinking the model’s memory footprint. […]
A new uncensored model is now available on Hugging Face and it carries the long name Qwen3.6-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking-NEO-CODE-Di-IMatrix-MAX-GGUF. This release is a 40-billion-parameter dense language model that was expanded from Qwen's […]
MiMo-V2.5-Pro-FP4-DFlash is a new release that combines expert-level 4-bit quantization with block-diffusion speculative decoding to shrink both the memory footprint and the number of forward passes needed during decoding. The […]
SCAIL-2 is a new open-source model that animates still character images directly from a driving video without relying on skeleton maps or inpainting masks. This end-to-end approach removes information loss […]
Qwopus3.6-27B-Coder-MTP-GGUF is a new quantized coding model designed for fast, local agentic software development. It represents a specialized version of the Qwopus3.6-27B-Coder, packaged as a GGUF file for efficient single-GPU […]
Gemma-4-12B-OBLITERATED is the first language model to completely remove built-in safety refusals with absolutely zero loss in benchmark performance. This modified version of Google’s Gemma 4 12B scores identically to […]
North-Mini-Code-1.0 is a new open-source AI model from CohereLabs built specifically for code generation, autonomous software engineering, and terminal tasks. It packs a 30-billion-parameter architecture but only activates 3 billion […]
MiniMax-M3 is a new native multimodal AI model from MiniMaxAI that processes text, images, and video with a 1 million token context window. The model contains about 428 billion total […]
Moonshot AI has released Kimi-K2.7-Code, an open-source coding agentic model that significantly upgrades long-horizon software engineering performance. It builds directly on the Kimi K2.6 architecture while cutting thinking-token usage by […]
Google has released a new open-weights AI model called Diffusiongemma-26B-A4B-it, which uses a unique method to generate text significantly faster than traditional models. Instead of creating text one token at […]
Anvil is a new open-source model runner that wraps llama.cpp, providing transparent controls and fleet management for local AI on private hardware. It stores models as plain GGUF files you […]
MindLab Research has released Macaron-V1-Preview-749B, a 749-billion-parameter AI model built to serve as a personal agent that can use tools and generate dynamic user interfaces. The system combines a massive […]
Vllm-doctor is a new command-line tool that diagnoses performance bottlenecks in vLLM inference servers by reading live metrics. It turns raw data into clear explanations of what is wrong, why […]
Meddies PII is a new open multilingual model that extracts personally identifiable information (PII) from clinical and administrative text, returning structured JSON instead of raw token tags. It identifies seven […]
A new release of AI Upscaling Models from Limitlesslab provides a set of custom-trained neural networks for photorealistic image and video enhancement. The collection includes GAN models built on architectures […]
Load images one frame at a time ComfyUI-SequentialImageLoader is a custom node that loads images from a folder one by one in order, advancing to the next file every time […]
ComfyUI-Workflow-Image-Export now includes experimental support for the newer ComfyUI Node 2.0 frontend. This extension saves a clean screenshot of your entire workflow and can embed the workflow JSON directly into […]
Comfyui-lrw-nodes is a newly released set of custom nodes for ComfyUI that brings latent-space geometry tools into local AI image and video workflows. It lets users experiment with geodesic interpolation, […]
NeuralCompanion is a new local Windows application that turns your PC into a realtime AI companion with chat, speech, avatar animation, and visual replies. It pulls together multiple services so […]
ComfyUI-ROCm-Windows-Native is a new open-source release that automates setting up ComfyUI to run directly on AMD GPUs in Windows using the native ROCm/HIP backend. It completely sidesteps Microsoft’s DirectML translation […]
AniGen-mac is a new port that brings the AniGen 3D asset generator to Apple Silicon Macs. It creates a fully rigged, animate-ready 3D model from a single input image. The […]
A new open-source tool called ideogrammar gives anyone a visual way to build structured prompts for the Ideogram 4 image model. It is a single web page that runs without […]
Google has released Magenta-Realtime-2, an open music generation model designed to create music on your own device with extremely low delay. This new system lets you steer musical output in […]
A new open-source project called Hitokudraft brings a fully local, voice-first AI assistant directly to the macOS menu bar. This native application runs entirely on Apple Silicon, processing voice commands […]
ProveKV is a new open-source release that slashes the memory footprint of multi-agent language model systems by up to 68 times. It stores the shared part of a conversation once […]
Horus-Lens-1.0 is an advanced text-to-image and image-to-image generation model that was extensively fine-tuned on a curated dataset of hundreds of thousands of clean images to improve prompt understanding and artistic […]
The Advanced-GGUF-Quantizer toolkit is a new CUDA-powered utility for building highly optimized GGUF models, with special attention to NVIDIA’s NVFP4 and MXFP6 data types. It uses a refined scale fitting […]
Supra-50M-Reasoning is an experimental open-source language model that produces a step-by-step thinking process before giving its final response. Designed as the reasoning version of Supra-50M-Instruct, it adds a chain-of-thought segment […]
Google has released Gemma-4-31B-it-qat-w4a16-ct, a compressed version of the new Gemma 4 31B instruction-tuned model that uses Quantization-Aware Training (QAT) to dramatically cut memory use while preserving high performance. The […]
A new open-source release called agent-sh introduces a composable agent runtime that embeds a lightweight coding assistant directly into your terminal. You can pair any frontend with any agent backend […]
Dots.tts is a new 2-billion-parameter text-to-speech model that converts text directly into high-fidelity 48 kHz audio without relying on discrete audio codec tokens. The system operates fully end-to-end, using an […]
OpenLumara is a new open-source AI agent framework built from the ground up in Python, designed to run locally and keep token counts low. It serves as a private, modular […]
Comfyui-navigator is a floating control panel that automatically lists every group in your ComfyUI workflow. It lets you jump to any group with a single click, toggle groups on or […]
The ComfyUI-DoRA-Dynamic-LoRA-Loader by Xmarre now includes a State Manager that lets users save and reuse entire character and style presets containing multiple LoRA stacks, prompt templates, and settings. The loader […]
Hildegard is a new ComfyUI node pack and LoRA for tiled upscaling and refinement on FLUX.2 Klein. It splits images into tiles, creates three context latents per tile, and blends […]
F2_k_Spectral_Graft is a new ComfyUI custom node that performs regional image editing on the Flux.2 Klein model without altering the rest of the scene. Users paint a mask over an […]
The Akium-sampler is a new momentum-guided diffusion sampler that fits right into Stable Diffusion Forge and ComfyUI. It’s a second-order stochastic sampler, meaning it adds controlled randomness at each step […]
The Comfyui-Lance-Aio release brings ByteDance’s Lance-3B model into ComfyUI with custom nodes that handle seven multimodal tasks. It can generate images and video from text, edit both media types, and […]
HuggingFace_WFX is a new file system plugin that turns Hugging Face model repositories, organizations, and collections into a virtual folder tree you can browse inside Total Commander on Windows. It […]
MisoTTS is a new 8 billion parameter text-to-speech model now available on Hugging Face. It converts written text into natural, conversational speech while maintaining voice consistency from short audio samples. […]
NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 is a new large language model that packs 550 billion total parameters while activating only 55 billion during use. It combines Mamba-2, mixture-of-experts (MoE), and attention layers into a […]
Boson AI has released higgs-audio-v3-tts-4b, a 4-billion-parameter text-to-speech model designed specifically for conversational voice AI. Rather than simply reading text aloud, the model produces expressive speech with emotional tone, natural […]
The Nemotron-3.5-ASR-Streaming-0.6b model is NVIDIA’s latest open speech recognition release, designed to transcribe audio in real time across 40 language-locales from a single model. It can handle both low-latency streaming […]
Eclipse-Senpai has released KeyLM-75M, a compact 75 million-parameter language model trained from scratch on roughly 18 billion tokens. This base text-completion model outputs plain text completions and is accompanied by […]
Hcompany has released Holo-3.1-0.8B, the smallest model in a fresh family of vision-language models built to drive computer use agents. The release expands automation capabilities beyond web browsers and desktops […]
Holo-3.1-35B-A3B is the largest model in a new family of vision-language agents that can see, understand, and control computer interfaces across web browsers, desktops, and now mobile devices. It automates […]
Nex-N2-mini is a new open-source AI model designed to handle complex, multi-step tasks by turning its own reasoning into real actions. It is the smaller, more efficient sibling of the […]
Nex-N2-Pro is a new open-source AI model designed to handle complex, real-world agentic tasks. It unifies reasoning, tool use, and code execution into a single continuous workflow called Agentic Thinking. […]
KVarN is a new KV-cache quantization method that expands context capacity for large language models. It compresses keys and values to 4-bit and 2-bit without calibration, yielding up to 3-5x […]
SDMLX is a new alpha-stage ComfyUI node suite that runs SDXL natively through Apple’s MLX framework on Macs. It converts a checkpoint into a .sdmlx package once, then loads it […]
A new set of custom nodes called ComfyUI_RaykoStudio is now available for the ComfyUI interface, focusing on interactive image manipulation. It bundles 15 distinct tools ranging from collage making and […]
HOT-Step-CPP is a new release that delivers a complete, locally-run AI music studio with a polished graphical interface. It wraps the acestep.cpp inference engine, letting you describe a song with […]
MOSS-SoundEffect-v2.0 is an open model that creates high-fidelity sound effects straight from text prompts. It can generate rain, city noise, animal calls, human actions, and even short musical clips with […]
Comfyui-Client-Android is a new native Android app that turns your phone or tablet into a direct controller for your ComfyUI image generation server. It eliminates the pain of trying to […]
ComfyUI-Image-Oasis is a new all-in-one node that packs an entire image generation workflow into a single, self-contained module for the popular AI art tool ComfyUI. Instead of stringing together loaders, […]
MBQ Viewer is a standalone desktop application that reads the embedded metadata inside ComfyUI PNG files and displays it in a clean, color-coded panel next to your images. It also […]
ComfyUI-FileManaty is a new custom node that embeds a full file manager directly into the ComfyUI web interface. You can browse, preview, rename, move, copy, upload, and even trash your […]
SPAG4d now offers the PaGeR generator, a depth-based backend that produces the most geometrically consistent 3D Gaussian splats from a single 360‑degree panorama. The tool turns any equirectangular photo into […]
JoyAI-Echo is a new open-source model that generates multi-shot, minute-long videos with synchronized audio directly from text prompts. You give it a JSON script describing each shot, and it outputs […]
The Damn-Simple-ComfyUI-Manager is a portable Windows app that lets you install, organize, and run multiple local ComfyUI instances from one clean interface. It removes the need to bounce between folders, […]
ComfyUI-KSampler-Matrix-Lab is a custom node for ComfyUI that generates a labeled comparison grid of different sampler and scheduler combinations. It runs each selected pair sequentially using the same base settings […]
Fizgig, a free open-source LoRA studio purpose-built for Flux 2 Klein 9B, has been updated to version 1.7 with a set of live control and reference-image tools. The release lets […]
ByteDance has released Bernini-R, the open-source renderer that powers its Bernini video generation and editing framework. The release provides inference code and model weights, letting users generate and edit videos […]
Cosmos3-Super is a new release from NVIDIA that generates video, images, audio, and even robot action plans from mixed inputs like text, photos, and video clips. It's an omnimodal world […]
The Ideogram-4-fp8 release is a new quantized version of Ideogram AI’s open-weight image generation model, using FP8 precision to run on a much wider range of hardware than its CUDA-only […]
JetBrains has released a new open-source AI model called Mellum2-12B-A2.5B-Thinking, a specialized reasoning tool that shows its work before giving you an answer. Unlike standard chatbots, this model wraps its […]
Google DeepMind recently released Gemma 4, and the Gemma-4-12B-It-GGUF version from Unsloth makes the 12 billion parameter model ready for local use. This quantized format shrinks the model files dramatically, […]
Google just released Gemma-4-12B-It, an open-weights instruction-tuned model that handles text, images, video, and audio natively in one compact 12 billion parameter package. Instead of bolting on separate vision and […]
Google has released Gemma-4-12B, a 12-billion-parameter open model that handles text, images, and audio in a single decoder-only system. The unified design ditches separate encoders, so all data goes straight […]
Tapping into the mathematical beauty of fractals, Openmandel is a new tool that lets AI models generate and explore detailed Mandelbrot images on command. It functions as an MCP server, […]
VibeETL is a self-hosted, visual ETL platform that lets you build data processing pipelines by dragging and dropping nodes onto an interactive canvas. Instead of writing code to move and […]
The Realtime-Multilingual-Asr-Router is a lightweight coordinator that routes live audio between small, specialized speech recognition models instead of relying on a single large multilingual system. It detects speech boundaries with […]
The Comfyui-preset-gallery is a new visual extension for ComfyUI that saves, organizes, and reuses your go‑to prompt snippets and style templates. It adds a dedicated Preset Gallery node, letting you […]
img2imgVideoTransparency is a new ComfyUI workflow that creates short video clips with transparent backgrounds from just two images. It uses Wan 2.2 image-to-video models, LightX2V LoRAs for fast generation, and […]
Eisbach-Medium is a new LoRA adapter for the Stable Audio 3 Medium model that generates long-form instrumental music with clear narrative development. The adapter weighs only 41MB and uses a […]
TripoSplat is a new open-source model that turns a single image into a 3D Gaussian splat, with a variable point count up to 262,144. It runs locally on a consumer […]
Flux_ID_Adjuster_V2 is a new ComfyUI node that improves identity consistency and realism in Flux.2 Klein 9B images. It tackles the common waxy skin problem by letting users control how reference […]
Cosmos3-Super-Text2Image is a new open-source image generation model from NVIDIA that creates high-fidelity pictures directly from written prompts. It is a text-to-image variant of the broader Cosmos3 platform, which is […]
Cosmos3-Super-Image2Video is a new open-source AI model that converts a single image into a short video clip guided by a text description. Released by NVIDIA, the 64-billion-parameter tool along with […]
ByteDance has released Bernini, an open-source framework that unifies video generation and editing through a semantic planning approach. Instead of controlling pixels directly, the system uses a multimodal large language […]
The Flux-2-Klein-9B-Schematic-Lora release offers a set of six LoRA adapters that reframe common computer vision tasks as simple image-editing jobs. Each adapter produces a schematic RGB output—like a depth map […]
Nvidia has released Cosmos3-Nano, a 16-billion-parameter omnimodal model that turns text, images, video, audio, or action data into dynamic video with synced sound, reasoning text, or robot movement commands. The […]
The new Bonsai-Image-Ternary-4B-Gemlite-2bit model compresses a 4-billion-parameter text-to-image diffusion transformer into just 1.21 GB. It uses ternary weights—each limited to -1, 0, or +1 with shared scaling—to shrink the model […]
The Comfyui-Anima-IPadapter custom node brings IP-Adapter support to the Anima DiT model inside ComfyUI. It lets users inject reference image features directly into the generation process using decoupled cross-attention. This […]
LongLive-RAG is an open-source framework that turns long video generation into a retrieval problem. An autoregressive generator can look back over its own output and pull in the most relevant […]
Prism ML has released Bonsai-Image-Binary-4B-Gemlite-1bit, a text-to-image model that packs a full diffusion transformer into just 0.93 GB by using binary weights. It takes the FLUX.2 Klein 4B architecture and […]
Autoswarm is a new open-source tool that turns local language models into self-improving assistants. It works as a drop-in API proxy that logs your chats and distills useful lessons into […]
Gemma-4-Harmonia-31B-uncensored-heretic is a decensored version of a 31-billion-parameter language model that dramatically cuts response refusals by 91%. The release uses an ablation technique to strip away content restrictions while keeping […]
PaddleOCR-VL-1.6 is a compact document parsing model that reaches a new state-of-the-art accuracy of 96.33% on the OmniDocBench v1.6 benchmark. It boosts recognition of text, formulas, tables, ancient documents, rare […]
Lance-2080ti is a hardware-optimized fork of ByteDance’s Lance model, tailored specifically for RTX 2080 Ti cards with 22GB of memory. It runs video and image generation without relying on BF16 […]
LFM2.5-8B-A1B is an open model that puts 8.3 billion total parameters into a package where only 1.5 billion activate per token, letting it run assistant workloads efficiently on local hardware. […]
VTS (Voice To Sound) is a newly released open-source model that turns a short vocal imitation and a text description into a realistic sound effect. Instead of fumbling to describe […]
Pantheon-Reasoning-27B is a new open-source language model that adds explicit step-by-step reasoning to roleplay and storytelling. It is built on an uncensored version of Qwen 3.6 27B and learns to […]
Odysseus 1.0 is a self-hosted AI workspace that gives you a ChatGPT-like experience running entirely on your own hardware. The tool brings chat, agents, document editing, and research capabilities into one […]
ComfyUI-AnimaFastTrain is a new set of custom nodes that lets you train temporary reference tokens for Anima models directly inside ComfyUI. These nodes work completely in memory without saving any […]
The ComfyUI-CNS-Sampler-CHENGOU is a new custom node that brings Colored Noise Sampling to ComfyUI, improving image generation by routing noise energy to the frequency bands that need it most. Unlike […]
A new open-source release called Colored-Noise-Sampling changes how diffusion models add randomness during image generation, giving better results without extra training. Standard samplers treat all frequency details the same and […]
NAVA is a new open-source model that creates synchronized video and audio from a single text prompt. It generates 720p clips with stereo sound, supports multiple speaking voices with distinct […]
Qwen-Image-Bench is a new evaluation toolkit that scores images from any text-to-image model using a fine-tuned judge AI called Q-Judger. It checks generated images across five major quality dimensions, including […]
Am-I-Openai-Compatible is a new command-line probe that tests any HTTP server claiming OpenAI API compatibility and shows exactly which endpoints work as promised. It runs a fast two-phase audit on […]
Equinox-31B is a new storytelling model from LatitudeGames that blends dark, high-stakes adventures with quiet slice-of-life narratives. It was fine-tuned on Google’s Gemma 4 31B Instruct using two distinct creative […]
The cohere-transcribe-diarize model adds speaker identification and word-level timestamps to Cohere’s open source speech recognition system. It works by fine-tuning the existing Cohere Transcribe model to output special tokens that […]
A new quantized version of Alibaba's coding model has been released to the community, offering a 27B parameter AI that can run entirely on a single 16GB graphics card. The […]
Llampart is a local-first chat web interface designed to pair with the llama-server backend. The release brings a polished desktop experience with a custom Frosted Glass visual theme. It lets […]
A new release called TTS-bench offers a practical way to compare local text-to-speech models across different hardware and listening tests without installing anything. The project provides a gallery of audio […]
TradingAgents-GUI gives anyone the ability to run a full multi-agent stock analysis locally through a polished web dashboard. It replaces the command line with a browser interface that walks you […]