LangChain v1 (LCEL + create_agent)

LangChain v1: exposes LLMs, prompts, tools, retrievers, and parsers as Runnables composed via the `|` operator (LCEL) into chains with uniform invoke / stream / batch semantics. create_agent returns a LangGraph CompiledStateGraph.

✓ 0 reported success·v0.1.0·

Overview

LangChain is the de-facto standard Python framework for building LLM applications (github.com/langchain-ai/langchain). The v1 package (v1.2.15) is intentionally minimal: core is `agents.create_agent` (returns a LangGraph CompiledStateGraph), `chat_models.init_chat_model` factory, message-type re-exports, and tools/embeddings shims. The historical `Chain` / `LLMChain` / `Memory` / `AgentExecutor` surface lives in `langchain-classic` (compatibility package). LCEL makes `prompt | llm | parser` pipe composition the standard pattern, covering ReAct agent loop, structured output, human-in-the-loop, model fallback, PII redaction, conversation summarization, TODO tracking, and RAG. This skill embeds 51 constraints covering typical pitfalls: BaseMemory ABC removed from langchain_core, LLMChain and Chain.__call__ all @deprecated, mismatched migration paths between AgentExecutor and create_*_agent, and provider partner packages requiring separate install. The host AI applies these automatically.

Blueprint Source

finance-bp-132

langchain-ai/langchain87ba30f1 source file

Constraints

1total
1fatal
1 must-not-violate

Evidence Quality

Confidence90%

High confidence — strong evidence base

1 non-negotiable constraints

FATALdomain_rulelangchain-C-007

WHENWhen defining a Pydantic model schema to pass to PydanticOutputParser, .with_structured_output(model), bind_tools, or any LCEL parser construction

ACTIONPick ONE pydantic version per parser/tool boundary and stick to it. Do NOT mix `from pydantic import BaseModel` and `from langchain.pydantic_v1 import BaseModel` (or `from pydantic.v1 import BaseModel`) in the same file. For new code use pydantic v2 (`from pydantic import BaseModel`). Migrating legacy v1 schemas: import is the discriminator — change ALL relevant model classes together, never partially.

CONSEQUENCEundefined behavior

FAQ

Discussion (0)

No comments yet. Be the first to share!

Changelog

v0.1.02026-04-25·Contributors: tangweigang-jpg

v0.1.0: Initial release on Doramagic.ai. LCEL framework on langchain-ai/langchain v1.2.15 with bilingual metadata, 51 anti-pattern constraints, and 3 FAQs.

v0.1.02026-04-25·Contributors: tangweigang-jpg

v0.1.0: Initial release on Doramagic.ai. LCEL framework on langchain-ai/langchain v1.2.15 with bilingual metadata, 51 anti-pattern constraints, and 3 FAQs.