AI Infra2025
Lets AI coding agents provision and operate a full backend themselves — Postgres with pgvector, OAuth2 auth, S3-style storage, Deno edge functions, and hosting — through one interface, plus an OpenAI-compatible model gateway.
Seven-week course that builds a production RAG system from scratch — an arXiv paper assistant that starts with BM25 keyword search, then layers hybrid vector retrieval, local-LLM generation, Langfuse monitoring, and an agentic LangGraph Telegram bot.
AI Agent2025
A TypeScript agent harness split into composable npm packages: a unified LLM API across OpenAI, Anthropic and Google, an agent runtime with tool calling and state, a self-extensible coding-agent CLI, and a differential-rendering terminal UI library.
AI Agent2025
Framework for building multi-modal AI agents that watch, listen, and reason over live video, pairing vision models (YOLO, Roboflow, Moondream) with LLMs like Gemini and OpenAI. Agents join calls in ~500ms and keep audio/video latency under 30ms.
AI Agent2025
Defines a predictable repository-level instruction file for coding agents, giving teams one place to document workflow rules instead of each tool inventing its own context format.
AI Agent2025
Makes the spec an executable artifact: you write intent in structured markdown and AI agents generate the plan, task breakdown, and code from it. A specify CLI and slash commands drive a constitution-plan-tasks-implement workflow across 30+ coding agents.
AI Agent2025
Wraps Claude Code in a loop that re-runs it until a task is done, gating every exit behind two conditions — semantic completion plus an explicit EXIT_SIGNAL — so it never stops early. Adds rate limiting and a circuit breaker for unattended, headless runs.
AI Agent2025
Composes AI agent teams from a Ghost+Shell+Model formula: each Bot pairs a prompt/MCP/Skills Ghost with a Chat, ClaudeCode, or Dify shell and a model like Claude or DeepSeek. Bots form Teams that run as traceable Tasks, wired to GitHub and DingTalk.
Teaches AI agent principles and practice through a structured Chinese curriculum, pairing theory with runnable code so learners can build, debug, and extend agent systems step by step.
AI Video2025
Generates explorable, 3D-consistent virtual worlds from a single image or short video. Includes official implementations of Lyra‑1 (feed‑forward 3D/4D scene generation via video-diffusion self-distillation) and Lyra‑2 (long-horizon, explorable generative 3D worlds). Best for research and creative prototyping; requires substantial GPU compute.
AI Infra2025
Extends vLLM beyond text to serve omni-modal models — Qwen3-Omni, TTS like CosyVoice3, and diffusion image/video/audio generators — in one engine, adding the non-autoregressive Diffusion Transformer support the core project never targeted.