Havuz

Run GLM-5.3-Flash (321B), DeepSeek-V4.1-Flash (748B) and MiniMax-M3.1-Flash (428B) on your own GPUs - one engine, one dashboard, OpenAI- and

Open decision models from 0.8B to 27B: a calibrated probability for every option in one pass, and a short thought only when unsure. Engine f

Run models larger than your RAM or VRAM by streaming MoE experts from SSD. Built on llama.cpp.

Atomic Agent, yerel veya cloud modellerle çalışan ve browser, dosya, onaylı shell komutları ile MCP araçlarını kullanabilen bir AI agent uygulamasıdır.

Neden yükseliyor · 159 günlük bir repoda 2.505 yıldız ve son 28 günde 648 commit dikkat çekiyor; yerel modellerle çalışan kapsamlı bir agent vaadi ve README'deki GAIA L1 karşılaştırması ilgiyi artırmış olabilir. Ancak commitlerin yalnızca 2 farklı yazardan gelmesi aktivitenin geniş bir katkıcı tabanına yayıldığını göstermiyor; yükselişin kesin nedeni bilinmiyor.

Rewrites AI drafts so they read like a person wrote them — runs locally, keeps the facts

Local inference for diffusion language models (LLaDA, Dream), written in Mojo

37 tok/s from Qwen3-Next-80B (48 GB) on a 32 GB laptop, 24 without drafts vs 2.8 for llama.cpp; reading 62 vs 16 tok/s. Faster than llama.cp

Screen Translator, Android ekranındaki yazıları OCR ile tanıyıp çeviren ve sonucu ekranın üzerine yerleştiren bir uygulamadır.

Neden yükseliyor · 97 günde 906 yıldız alması, özellikle oyun ve manga çevirisi gibi belirgin bir kullanım alanına hitap ettiğine işaret ediyor. Ancak son 28 günde 2 commit ve 1 farklı yazar bilgisi, ilginin sürdürülebilir geliştirme faaliyetiyle desteklendiğine dair sınırlı kanıt sunuyor.

Local AI app and inference engine for agents. Run open-weight LLMs locally — private, 100% offline on your computer. Join our Discord: https

Run MoE models bigger than your RAM. Frontier-size MoE on a phone, laptop, edge. Lossless, on stock llama.cpp. An adaptive streaming optimiz

Uncensored DeepSeek-V4.1-Flash IQ2_XS on Google Colab

Pure-Rust inference engine for Qwen GGUF models (Qwen3 to Qwen3.8, dense and MoE up to 80B+). CUDA and Vulkan (NVIDIA, AMD), hybrid CPU+GPU

llama.cpp fork for significantly improved performance on Ampere generation Nvidia cards (especially RTX 3090 / 3090 Ti): example: 110+ tok/s

A native .NET LLM inference engine and agent runtime for GGUF models. TensorSharp provides a console application, a web-based chatbot interf

A dependency-free C inference engine for Liquid AI's LFM2 / LFM2.5 models. Terminal chat, an OpenAI-compatible server, a web agent and a ter

Entrpi/ds4, a Blackwell CUDA perf fork of antirez/ds4 on NVIDIA DGX Spark: one-command install, ~3x upstream prefill, ~1.5x decode, DSpark,

Local decision model with calibrated probabilities: send a state and yes/no, choice or score questions, get a probability for every option.

High-performance full-text search across PDFs, Office documents, ZIP archives, email, media metadata and GGUF/LLM model files. ripgrep-compa

SSD-streaming inference engine for giant MoE models (Rust + CUDA). GLM 5.2 743B at 2 tok/s and Hy3 295B at 7 tok/s on two consumer 16GB GPUs

Run any local LLM engine, auto-tuned to your GPU — polished web UI + OpenAI/Anthropic-compatible API. Point Claude Code at your own machine

LLM inference with 7x longer context. Pure C, zero dependencies. Lossless KV cache compression + single-header library.

Zero-overhead, terminal-native local-LLM orchestrator. Run any engine, any model via the same proxy.

Evidence-backed AMD Strix Halo local-AI setup and benchmarks: Qwen3.8, Ollama, llama.cpp, Vulkan/ROCm, large GGUFs, and cross-OEM results.

A tiny neural-network framework in pure Go with AVX2 SIMD kernels (GOEXPERIMENT=simd)

Desktop workspace for quantizing LLMs to GGUF, MLX and ONNX. Watch every tensor change, compare against the original, and publish to Hugging

Look inside any local LLM (GGUF from Ollama, LM Studio, Hugging Face, Jan, llama.cpp): layers, heads, vocabulary, chat template, weights. Ex

A 6.87 GiB GGUF build of Qwen 3.8 27B — full 27B quality on 8 GB RAM. 96.7% HumanEval, 92.9% MMLU retention, no distillation.

Hybrid GPU + CPU inference for mixture-of-experts models, written in Rust down to the CUDA kernels. A drop-in llama-server.

Native AI assistant for macOS, built with SwiftUI and llama.cpp. Run local LLMs offline, work with PDFs, and create custom AI assistants. Op

llama.cpp/ik_llama.cpp launcher: loads big MoE models across mismatched multi-GPU rigs by exact VRAM math.

Camelid: a Rust-native local inference backend with evidence-gated model compatibility.

WHIRL (Windows HIP Inference for RDNA LLMs): native Windows C++/HIP LLM inference engine for the AMD Ryzen AI Max+ 395 (Radeon 8060S) and Ra

Private, offline AI assistant for Android and Windows. Run local LLMs (Gemma 4, Qwen, GGUF) on your phone or PC with llama.cpp, CUDA and Vul

JANG — GGUF for MLX. YOU MUST USE JANG_Q RUNTIME. Adaptive Mixed-Precision Quantization + Runtime for Apple Silicon

Open-source control panel for the real llama.cpp: see where a model lands in your VRAM before loading, every llama-server flag per model, on

ENGRAFT: write facts into an LLM's n-gram (Engram) memory table as a removable overlay: 100 facts in a 9 MB file, no weights touched, verifi

Hemmingway-1: a 27B model that writes the way a person writes. Open weights, free for non-commercial use (CC BY-NC 4.0); commercial use by a

Open, calibrated System 1 decision models: typed choice / score / yes-no answers with a probability for every option. 400M to 80B; CPU, NVID

CPU-first LLM inference server on llama.cpp. Runs useful models on free-tier ARM boxes; rewriting the input made it ~6x faster and more accu

Open-source desktop AI agent: built-in free model, local GGUF models via llama.cpp, dispatch tasks from WeChat/Feishu/DingTalk. Files, brows

Get more intelligence from every bit. Better quantization formats and smarter calibration let larger, stronger models run smoothly on the ha

Run a 110B on a 2016 PC with 16 GB RAM. Know your tok/s before you download. Placement beats budget: predicts speed + memory fit for any GGU

🦀 Decoder-only LLM built from scratch in pure Rust using Candle — no Python, no PyTorch. Gated DeltaNet + sparse attention, fine-grained MoE

A local LLM server built for concurrent work. Drop-in replacement for Ollama and/or OpenAI and Ollama APIs on one port. Requests that share

Gumbel-Softmax post-training quantization for LLMs (1–3 bit scalar, INT/GGUF-compatible).

Local-first AI chat client combining SillyTavern-style character cards, LM Studio-style GGUF model management, Ollama/OpenAI-compatible APIs

Run stock llama.cpp with the stock Unsloth Qwen3.6-35B-A3B 4-bit on a 16 GB Mac.

User friendly GUI for configuring and launching llama.cpp

Fine-tune Qwen2.5-Coder-7B for free (QLoRA on Kaggle T4) and run a local coding agent that runs your code instead of guessing — 8 GB Mac, ll

Ollama-compatible LLM server on a tuned llama.cpp fork. Beats vllm by 40% on single stream decode, and by 20% on multi stream decode, and be