Where AI Experience Lives
Infrastructure for the AI era. Manage your skills, MCPs, prompts and scripts in one place — and share them so any human or agent can install with a single command.
Standalone gems — hand-picked, refreshed every 4 hours
Author just shipped an update — back on the radar.
Author just shipped an update — back on the radar.
Author just shipped an update — back on the radar.
Author just shipped an update — back on the radar.
Author just shipped an update — back on the radar.
LibreTranslate (self-hosted DeepL alternative), Tolgee for code i18n, Weblate for community translation, plus PDF and video translation pipelines. Five no-rate-limit tools.
- ▸LibreTranslate — Self-Hosted Translation API with No Rate Limits
- ▸Tolgee — Developer-Friendly Localization Platform
- ▸Weblate — Web-Based Continuous Localization Platform
Chroma, Weaviate, Pinecone, txtai, Qdrant MCP, plus the embedding APIs from Cohere and Together — pick by latency, cost, or RAG accuracy.
- ▸Chroma — Open-Source Vector Database for AI
- ▸Weaviate — Open-Source Vector Database at Scale
- ▸Pinecone — Managed Vector Database for Production AI
Mem0, Zep, Cognee, and the patterns to make agents remember across sessions — without baking everything into the prompt.
- ▸Mem0 — Memory Layer for AI Applications
- ▸Codebase Memory MCP — Code Intelligence for AI Agents
- ▸Memory MCP — Persistent AI Agent Knowledge Graph
Seven Python and TypeScript agent frameworks side by side — CrewAI multi-agent, LangGraph stateful graphs, Phidata's tool-native runtime, Agno's reasoning, OpenAI Swarm, Pydantic AI, Mastra TS.
- ▸CrewAI — Multi-Agent Orchestration in Python
- ▸CrewAI Flows — Event-Driven Multi-Agent Orchestration
- ▸LangGraph — Stateful AI Agent Graphs by LangChain
Cursor (W22), Cognition (S23), Together AI (W22), Pinecone (W21), Vapi (W22), Replit (S09) — the YC-funded AI products that turned demo decks into production tools.
- ▸Awesome Cursor Rules — AI Coding Standards for 30+ Stacks
- ▸Cursor Rules MDC Generator — Auto-Generate from Docs
- ▸Devin Cursorrules — Turn Cursor into an AI Engineer
Deepgram Nova-3 (60ms partial), AssemblyAI Universal-2 (99 langs), Whisper on Groq (166× realtime) — the STT engines voice agents and call centers run on.
- ▸Deepgram Nova-3 — Production STT with 60ms Partial Latency
- ▸AssemblyAI Universal-2 — Streaming STT for Voice Agents
- ▸AssemblyAI Diarization — Auto-Identify 2-10 Speakers