Search

A webpack plugin to protect your generated bundle using Jscrambler Code Integrity

Create a new Yugami frontend — Next.js 16 + Tailwind v4 + @yugami/ui + RTK Query + Vitest. Single command: pnpm create @yugami/interaction

OpenAI-compatible provider for a Broapp application's AI layer: OpenAI, Ollama, LM Studio, vLLM and anything speaking the same API

nj41o4

The durable conversation record for AI SDK agents: provenance-cited derive-at-write append, ghost-guard reply extraction, turn fencing, fail-loud reads.

DSH web plugin: a LiveBench tab in the Trajectory view (right of 对话/轨迹). Run LiveBench evaluations against every model configured in the DeepSeek Harness — pick provider/model, category, task, release and question range from dropdowns, watch progress, and

Vite plugin that bundles only the Iconify icons your source actually uses — no runtime API calls, no whole collections.

Single source of truth for chain configs, token lists, logos, ABIs, and shared utilities across spiceflow repos.

An enterprise-grade, isomorphic, type-safe HTTP client tailored for Next.js App Router, Server Actions, and Client Components.

Short messages from other developers shown in your Claude Code spinner while the AI thinks — zero tokens, never reads your code. Claude Code / Codex integration

AppPort command line interface.

A local-first workspace for living worlds.

Record Playwright steps as a SlideMux clip bundle, with CLI and MCP.

Autonomous browser automation CLI tool to delete X (Twitter) posts, replies, reposts, unfollow accounts, and remove followers safely.

OQM (Optimized Quality Model): High-density suitcase squeezing, 5-bit packing, and on-demand layer-streaming LLM engine

Scan and safely clean local caches left by AI coding tools (Claude Code, Cursor, Ollama, npm/uv, HuggingFace, etc)

Z.ai (Zhipu) LLM client for AI-Parrot (parrot.clients.zai)

AgentRouter 中转聚合：一条 pi-ai 路由承载多模型 + 国内/国际端点一键切换（设置卡片）+ 出站请求 User-Agent/403-配额改写围栏。aqiu817/dsh-llm-agentrouter 分叉，适配 0.1.2-alpha.5 起新 settings/slots API，已在 0.1.3-alpha.2 逐项实证兼容。

GP regression + BALD active-learning engine over text embeddings, shared by OpenOutFind and OpenOutNews.

100M-parameter Transformer comparing 5-Component Attention (GAU) and MHA, with SmolLM tokenizer, Cut-Cross-Entropy, SwiGLU, and RMSNorm