GPT Image 2.5 Sunburst API (`gpt-image-2.5-ext`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
Seedance 1.5 Pro prompts: tested image prompt recipes with example renders, aspect ratios, the model id (`seedance-1.5-pro`) and the cost per image.
MiniMax H3 prompts: tested image prompt recipes with example renders, aspect ratios, the model id (`MiniMax-H3-Regeneration`) and the cost per image.
Kling 2.6 Motion Control prompts: tested image prompt recipes with example renders, aspect ratios, the model id (`kling-v2-6-motion-control`) and the cost per image.
FLUX Kontext API (`flux-kontext-pro`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
ハイルオ — MiniMax Hailuo 2.3 prompts: tested image prompt recipes with example renders, aspect ratios, the model id (`MiniMax-Hailuo-2.3`) and the cost per image.
MiniMax Hailuo 2.3 API (`MiniMax-Hailuo-2.3`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
Сидрим — Seedream 4.0 API (`seedream-4-0`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
God of War: Betrayal, remastered for Windows. A native PC port of the 2007 J2ME phone game with widescreen, high frame rates and gamepad support. Bring your own .jar.
Nano Banana 2 Lite API (`gemini-3.1-flash-lite-image-ext`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
FLUX.2 Flex prompts: tested image prompt recipes with example renders, aspect ratios, the model id (`flux-2-flex`) and the cost per image.
You describe your services once. Enodia handles the rest.
Steering files that configure an AI coding assistant to perform defensive evidence-based security vulnerability triage. Companion sample for the blog post "Configuring Your AI Vulnerability Harness: The Steering File."
—
ヴェオ / ベオ — Veo 3.1 Fast prompts: tested image prompt recipes with example renders, aspect ratios, the model id (`veo3.1-fast`) and the cost per image.
Grok Imagine Image Quality prompts: tested image prompt recipes with example renders, aspect ratios, the model id (`grok-imagine-image-quality`) and the cost per image.
Agent Deck: track Claude Code, Codex, WorkBuddy and DeepSeek Harness from Obsidian, delegate from the task line, and keep the agent summary in the macOS menu bar.
An alternative to the old Samsung DeX software on Windows.
The Official repository for the A4T-C tool changer project based on the A4T.
Talk風Live翻訳 — translated by PLaMo. 英語の動画を、読みやすい日本語のトークに。Windows向けのローカルLive翻訳アプリ。
—
—
FLUX.1 Kontext Max API (`flux-kontext-max`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
🐜 ant is a Golang ORM, that leverages modern generics to enable strict typing and incredible developer experience. Inspired by ASP.NET's Entity Framework and Laravel
Clean, config-driven LLM Post-Training: SFT, DPO, GRPO, OPD, SDFT - one YAML, any topology.
MCP server giving AI agents complete Tehran Stock Exchange (TSETMC) and Codal data: quotes, order book, tick trades, real/legal flow, shareholders, filings. بورس تهران برای ایجنتهای هوش مصنوعی
ALL IN 你的手机移动api网关
Cross-version Go pclntab auto-locator & patcher to unlock full 1M context window (983K) and 1MiB/128K AGENTS.md rules budget for Antigravity IDE, Desktop & agy CLI (Windows PE & Linux ELF).
agentic work organized visually.
ROC OS:目标为全 H700 共用的 Android 10。DisplayAuto Test2 因实机显示回归已拒绝,停止推荐;当前无新的显示验收通过包。 / Android 10 for H700. DisplayAuto Test2 rejected after physical display regression; no longer recommended. See README.
Ocarina of Time — PC Native Build. Website moved to ocarina-pc-native-build.github.io. Development progress and reported test results; no playable release yet.
Gran Turismo 5: unofficial PC port (static recompilation), split from the GT5P repo gt5 branch
A 10-week, staff-level curriculum and engineering roadmap focused on LLM inference systems, serving physics, custom kernels, disaggregated architectures, and production optimization.
Open-source AI-assisted PCB design for KiCad 10. Connect your AI assistant to schematic capture, PCB layout, routing, design checks, and manufacturing exports through MCP.
Windows Reimagined
会译:Windows 上的 AI 会议助手——实时字幕与翻译、发言人区分、AI 纪要、同声传译
Qwen3.8-Flash-Next on any consumer hardware. Strata inference engine, but without the bloatware. Designed with containerized environments in mind. OpenAI/Anthropic API on localhost, optional image input.
Studio-grade UI design systems, Design DNA extractor & Anti-Slop MCP server for AI coding agents (Claude Code, Codex, Cursor, Windsurf).
—
Cinematic interactive Earth-to-Mars mission simulator
Run Qwen3.8-Flash-Next on one AMD Strix Halo (Ryzen AI Max+ 395, gfx1151). EXL3 4.05 and 3.05 bpw packs, own ROCm/HIP kernels, OpenAI-compatible server, 4 sessions of 262K context, setup guide included.
HappyHorse 1.0 prompts: tested image prompt recipes with example renders, aspect ratios, the model id (`happyhorse-1.0`) and the cost per image.
A welcoming open-source creative studio for images, video and 3D, powered by ComfyUI. Make, edit and explore your ideas locally on your GPU or online.
Pokémon SoulLinks for 2–4 players and solo Nuzlockes on macOS, Windows and Linux – emulator from Game Boy to 3DS, live tracking, randomizer, real-time sync and OBS overlays in one app.
a compatibility and recompilation project that aims to bring PSP software to the PlayStation 2 by translating Allegrex code to the R5900 architecture and providing the runtime support needed for psp apis, graphics, audio, and system services
—
RideTube: an experimental, unofficial video client for the BMW iDrive 7
A self-hosted webmail client that runs entirely on Cloudflare Pages and Workers
Wukong(悟空)— 面向逆向分析与 Pwn 的多 Agent 工作台
—
Work in progress
Exact LLM inference engine written from scratch for AMD Strix Halo (gfx1151, ROCm/HIP). Qwen3.8-Flash-Next on one node or two replicas; DeepSeek V4.1 Flash across two nodes over USB4 with an optional RTX PRO. Speculative decoding bit-identical to greedy, up to 255 tok/s on coding chat. MiniMax and GLM next.
HappyHorse 1.0 API (`happyhorse-1.0`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
Hardware-aware MoE inference control plane for expert prediction, residency, and intelligent prefetching.
Make every Claude token count. A Claude Code plugin that picks the model and effort for every prompt.
A native file manager for Wayland: fast, light and safe. No GTK, no Qt, no GLib: C++20 on Wayland and OpenGL ES.
Tactile PWA Mobile Remote Deck & Universal AI Companion for Omarchy / Hyprland Linux
Mac storage cleaner/investigator and uninstaller
—
Zusia, a study buddy inside Zotero: ask Claude Code, Codex or Antigravity about the paper you are reading. LaTeX, drawings, theorem links, glass UI.