Modular RAG experimentation platform with FastAPI, Streamlit, multiple vector databases, reranking, and evaluation logging.
面向生产环境的 AI Agent 动态技能网关与自进化底座：三层级联意图路由(R@1=98%)、8维沙箱评估、8重防御防线、LangGraph状态图旁路与SQLite原子发布状态机。
HyDRA is a hierarchical deep research agent framework that orchestrates planning, tool usage, web exploration, and synthesis through robust context management, enabling reliable multi-step reasoning under strict output and token constraints.
Markdown document creation engine plus a ReAct visual layout optimizer. It renders synthetic Markdown documents to PDF, rasterizes PDF pages to PNG screenshots, computes deterministic layout metrics, and records human-in-the-loop ratings.
Autonomous multi-agent marketing pipeline built with LangGraph — takes a brand brief and outputs a fully researched, strategized, written, QA-reviewed, and analytics-forecasted campaign. Features ReAct agents, isolated tool registries, Pydantic-validated I/O, persistent brand memory, and LLM-agnostic provider switching (OpenAI · Anthropic · Llama).
A local-first autonomous LLM agent that perceives and navigates a Blazor WebAssembly UI using a ReAct loop, powered by Ollama, zero data leaves the machine.
To associate your repository with the re-act topic, visit your repo's landing page and select "manage topics."