Indras-Mirror / Xolotl

Xolotl (formerly Strata-DSV4.1): DeepSeek-V4.1-Flash, all 552B, on one gaming PC (RTX 4090 + 90 GB RAM) at ~27 tok/s, plus GLM-5.3-Flash, Qwen3.8-Flash-Next and Qwen3.8-27B. A heavily extended fork of Strata: MoE experts tiered across VRAM, pinned RAM and NVMe; OpenAI + Anthropic API.

NODIX-TECH / DLSS-5-MANAGER

THE FIRST DLSS 5 MANAGER

sudoingX / local-ai-on-your-gaming-gpu

run a 27b local ai on your gaming gpu, rtx 3060 and up, radeon rx 6600 and up in beta, 8gb or more, linux and windows, one line to install

professorpalmer / bonsai-ada-surgery

Bonsai 2 27B: 262k context (trained max) on 12 GB, 54 to 97 tok/s decode, 2x prefill, q8_0 KV recipe, grammar tool calls 9/9. Same ternary weights. Windows bundle + receipts.

2600th / dlss5-video-player

Windows player that runs video, photos or GIFs through NVIDIA's DLSS 5 neural renderer, then compares the result with the original frame by frame. Local files or YouTube. Needs an RTX card and driver 610.47+. Community project, unsigned runtime. C++20, D3D12.

ImDreamt / MFGAdaUnlock-RenoDx

DLSS MFG on Ada — unlocks 3x/4x/6x frame generation on RTX 40-series cards and corrects the temporal midpoint. In-memory only.

rashid2280bd / Optiscaler-Client-V4-Pro-2026

OptiScaler — Open-source graphics middleware allowing users to swap, configure, and enable DLSS, XeSS, FSR, and Frame Generation technologies across various GPUs in modern PC games.

alphastorm / omp-ninfer

Durable local inference for Oh My Pi: NInfer + Qwen3.8 27B on one RTX 3090/4090/5090, with restart-resumable OpenAI Responses state.

sepiablue-ai / ComfyUI-H3-X2-Stream

ComfyUI nodes for MiniMax H3: INT8 X2 VAE and asynchronous NVENC saving. 5-second 1088x1920 video with audio in about 60 s on an RTX 4070 12GB, about 30% faster than the FP16 X2 VAE.

xilla420 / dlss5-wow-sidecar

DLSS 5 Neural Rendering for World of Warcraft (WoW) on NVIDIA RTX 40/50 — a safe overlay with no DLL injection, no hooks and no ReShade next to Wow.exe. Free and open source.