把 PDF、DOCX、HTML 等文档解析成结构化 Markdown 或 JSON，方便喂给大模型和 RAG 流程。
给 AI Agent 做强化学习训练和调优的框架，把 Agent 的推理轨迹变成训练信号，让智能体越跑越强。
Changes since langchain-core==1.7.6 release(core): 1.7.7 ( 41776) fix(core): cap nesting depth in load() ( 41742) chore(deps): patch newer-python security lock resolution
Changes since langchain-fireworks==1.7.2 release(fireworks): 1.7.3 ( 41773) fix(fireworks): default max tokens from model profiles ( 41772) chore(model-profiles): refresh
- Preserve LangGraph Store key escaping, empty-index handling and backend write/list error handling ( 7038) - Preserve Unicode archive query constraints in text
Changes since langchain-typesafe==0.0.1 release(typesafe): 0.0.2 ( 41720) feat(typesafe): accept a classifier in experimental middleware ( 41657)
Changes since langchain-core==1.7.5 release(core): 1.7.6 ( 41698) perf(core): merge streamed AIMessageChunk s once when accumulating runnable streams ( 41697) perf(core):
Changes since langchain-openai==1.7.3 release(openai): 1.7.4 ( 41683) chore(deps): require langchain-core =1.7.5 for refusal handling ( 41682) feat(core): standardize mod