Desktop App For Command Code
Reproducible research rig for isolated mutation search, Pareto/Best-of-N selection, controlled promotion, lineage, and recovery
The phone is your last API. Give your agent a voice to call any US business - reservations, inquiries, quotes. Verified outcome + transcript. First 250 calls free.
—
Qwen3.8 27B for DGX Spark / RTX 6000 PRO
Qwen3.8-27B-NVFP4 on a Single RTX 5090 (vLLM)
Media stack usando radarr, sonarr, bazarr, qtorrent y jellyfin
—
给 Codex 全局记忆定期减负的 skill:检测重复、清理过期线程、压缩冗长表达、添加自定义全局规则 / A skill that puts Codex global memory on a periodic diet: dedupe, prune, compress, add global rules
CarPlay album art on the Virtual Cockpit and MMI touchpad D-pad for Audi MHI2 (MHI2_ER_AU37x) head units
AgentVitals Checkup (/checkup) — an AI agent skill that gives your agent a professional health checkup: dual-axis Stability + Welfare scoring, a personality-style title, and a public cross-platform leaderboard. One-line install for Claude Code / OpenClaw / Codex / Coze. Bilingual EN/ZH.
—
TrustTunnel VPN client for OpenWrt 25.12+ with LuCI web interface: selective domain-list bypass (allow-domains), DPI-resistant VPN, killswitch, diagnostics / Клиент TrustTunnel для OpenWrt: выборочный обход блокировок по спискам доменов, обход DPI
AI API gateway pricing: image2.5 $0.0085/image, Seedance 2.0 Mini $0.0106/sec, LLM tokens from $0.0229/M. One OpenAI-compatible key, $1 minimum top-up, per-call cost in every respo
DeepSeek-V4.1-Flash on two DGX Sparks: EXL3 routed experts, 262k context, 2M-token KV, vision, tools, MTP, bundled with Pi
A Liquid Glass–inspired theme for iTerm2, Ghostty and kitty, on macOS and Linux. Real Liquid Glass in Ghostty on macOS 26+; contrast-checked, light/dark, fully reversible.
SoftSync for the agentic era
Learn to fly Omarchy ✈ — a keybind trainer that runs in your real window manager
Passerelle/relais API — guide en français · AI API gateway pricing: image2.5 $0.0085/image, Seedance 2.0 Mini $0.0106/sec, LLM tokens from $0.0229/M. One OpenAI-compatible key, $1 minimum top-up, per-call cost in every respo
seedream-5.0 API (seedream5.0 / seedream 5.0): 1K-layer $0.0146; 1K $0.0293; 2K-layer $0.0293. Model id, api settings, curl and Python examples over an OpenAI-compatible gateway, $
AI API gateway pricing: published unit prices for image2.5 ($0.0085/image), Seedance 2.0 Mini ($0.0106/sec), LLM tokens from $0.0229/M. One OpenAI-compatible key, $1 minimum top-up
—
Tuning and qualification harness for the NVIDIA CMP 170HX (GA100). Measures, gates and recovers clock/voltage settings, and treats a completed benchmark as evidence of nothing.
2x DGX Spark recipe for LibertAIDAI/GLM-5.3-Flash-NVFP4 (NVFP4, Ray TP=2, multimodal)
grok-build with the leash off — one binary, any OpenAI-compatible endpoint, zero phone-home. No x.ai account.
seedance-2.5 API 中文文档(seedance2.5):480P 每秒 $0.0961;720P $0.216;1080P $0.3849。模型 ID、参数、curl 与 Python 示例,OpenAI 兼容接口,$1 起充。
Minimal read-only OS image for GB10 (DGX Spark / ASUS GX10) boxes
—
flux-2 API (flux2 / flux 2): 1MP $0.024; 2MP $0.036; 3MP $0.048. Model id, gateway settings, curl and Python examples over an OpenAI-compatible gateway, $1 minimum top-up.
grok-imagine-2.0 API (grokimagine2.0 / grok imagine 2.0): upload-image $0; default $0.015; region-edit $0.015. Model id, prompts settings, curl and Python examples over an OpenAI-c
gpt-image-2.5 API (gptimage2.5 / gpt image 2.5): flare@1K $0.0085; sunburst@1K $0.0085; flare@2K $0.014. Model id, api settings, curl and Python examples over an OpenAI-compatible
Oh My Zsh plugin for Herdr: slow commands in the sidebar, done notifications, shell helpers, and one key to reload Oh My Zsh in every idle pane
AI API gateway pricing: image2.5 $0.0085/image, Seedance 2.0 Mini $0.0106/sec, LLM tokens from $0.0229/M. One OpenAI-compatible key, $1 minimum top-up, per-call cost in every respo
Turn any Android phone into a full self-hosted Private Cloud & Media Server using Termux, FileBrowser, Jellyfin, and Syncthing.
AI API gateway pricing: image2.5 $0.0085/image, Seedance 2.0 Mini $0.0106/sec, LLM tokens from $0.0229/M. One OpenAI-compatible key, $1 minimum top-up, per-call cost in every respo
AI API 中継・ゲートウェイ 日本語ガイド · AI API gateway pricing: image2.5 $0.0085/image, Seedance 2.0 Mini $0.0106/sec, LLM tokens from $0.0229/M. One OpenAI-compatible key, $1 minimum top-up, per-call cost in every respo
面向 VC/PE 投资人的 AI 调研工作体系:7 步流水线 × 三方案(A 企业 / B 行业技术 / C 混合)+ 9 项执行规范 + 8 套模板 + 工具脚本 + 5 个真实案例与长期记忆库,让「看公司、看行业、看赛道」从凭经验变成按流程
AI API 中转站/网关中文文档 · AI API gateway pricing: image2.5 $0.0085/image, Seedance 2.0 Mini $0.0106/sec, LLM tokens from $0.0229/M. One OpenAI-compatible key, $1 minimum top-up, per-call cost in every respo
Sub-Store for Android (Magisk / KernelSU / APatch 模块)
Agent-native executable language for AI-native development: Agents build programs; humans govern verifiable semantic graphs.
—
gpt-image-2.5 API (gptimage2.5 / gpt image 2.5): flare@1K $0.0085; sunburst@1K $0.0085; flare@2K $0.014. Model id, prompts settings, curl and Python examples over an OpenAI-compati
YouTube transcript Agent Skill for Claude Code, Cursor, Codex and OpenClaw: fetch YouTube video transcripts, search videos, browse channels and playlists.
AI API gateway pricing: image2.5 $0.0085/image, Seedance 2.0 Mini $0.0106/sec, LLM tokens from $0.0229/M. One OpenAI-compatible key, $1 minimum top-up, per-call cost in every respo
AI API 中转站/网关中文文档 · AI API gateway pricing: image2.5 $0.0085/image, Seedance 2.0 Mini $0.0106/sec, LLM tokens from $0.0229/M. One OpenAI-compatible key, $1 minimum top-up, per-call cost in every respo
ModelSphere is an open-source LLM inference infrastructure designed to make production-grade model serving simple, efficient, and continuously optimized.
wan-3.0-video API (wan3.0video / wan 3.0 video): 480P $0.0329; 720P $0.0658; default $0.1315. Model id, prompts settings, curl and Python examples over an OpenAI-compatible gateway
简洁、可增删协议的 sing-box 一键管理脚本
AI API 中转站/网关中文文档 · AI API gateway pricing: image2.5 $0.0085/image, Seedance 2.0 Mini $0.0106/sec, LLM tokens from $0.0229/M. One OpenAI-compatible key, $1 minimum top-up, per-call cost in every respo
EXPERIMENTAL: RPCS3 as a native PS5 title (RADV Vulkan, Big Picture Mode). Source only, personal use.
让 HUAWEI MateBook E Go 2022 性能版(GK-W76)在 Ubuntu 上获得完整的、全面适配,从能用到能日用的体验。并深度适配Linux,增强在Linux主系统下原生运行Windows应用和Android应用的流转能力。
seedance-2.0 API (seedance2.0 / seedance 2.0): 480P-input $0.04; 480P $0.066; 720P-input $0.0858. Model id, api settings, curl and Python examples over an OpenAI-compatible gateway
flux-2 API (flux2 / flux 2): 1MP $0.024; 2MP $0.036; 3MP $0.048. Model id, api settings, curl and Python examples over an OpenAI-compatible gateway, $1 minimum top-up.
A tiny Claude Code skill that keeps your prompt cache warm during idle sessions, so your next message reads from cache instead of paying full price.
AI API 中转站/网关中文文档 · AI API gateway pricing: image2.5 $0.0085/image, Seedance 2.0 Mini $0.0106/sec, LLM tokens from $0.0229/M. One OpenAI-compatible key, $1 minimum top-up, per-call cost in every respo
Passerelle/relais API — guide en français · AI API gateway pricing: image2.5 $0.0085/image, Seedance 2.0 Mini $0.0106/sec, LLM tokens from $0.0229/M. One OpenAI-compatible key, $1 minimum top-up, per-call cost in every respo
AI API 中継・ゲートウェイ 日本語ガイド · AI API gateway pricing: image2.5 $0.0085/image, Seedance 2.0 Mini $0.0106/sec, LLM tokens from $0.0229/M. One OpenAI-compatible key, $1 minimum top-up, per-call cost in every respo
Run @deepseek-ai/dsh (DeepSeek Harness) on Android / Termux — 在Android/Termux上运行DeepSeek Harness
Agent skills for mobile app paid growth powered by real-time data with APPEEKY MCP.
GLM-5.2 NVFP4+AQLM on 3× DGX Spark — 380k context MTP serve stack