langserve

What it is and what it does

LangServe is a library for exposing LangChain runnables and chains as REST APIs. It uses pydantic for automatic input/output schema inference and validation. The library provides `/invoke`, `/batch`, and `/stream` endpoints for efficient concurrent request handling, plus a `/stream_log` endpoint for intermediate step visibility and a web playground for testing.

The package is designed for simple runnables and well-known LangChain primitives. It includes a Python client SDK to call deployed runnables as if they were local objects, optional tracing integration with LangSmith, and automatic API documentation. However, the repository is now archived and no longer accepting new features—the maintainers recommend LangGraph Platform for new projects.

Use it for:

• Deploy a LangChain LLM chain as a REST API with automatic schema validation and documentation.
• Expose a retrieval-augmented generation pipeline as streaming endpoints for concurrent requests.
• Build a conversational agent with message history and expose it via HTTP for web clients.
• Stream intermediate reasoning steps from a chain or agent to a frontend or monitoring system.
• Call a deployed LangChain runnable from a Python client using httpx or orjson serialization.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

LangServe deploys LangChain runnables and chains as REST APIs with automatic schema inference, data validation, and streaming support.

No—not for new projects. The repository is archived and abandoned; maintainers explicitly recommend LangGraph Platform instead. Existing LangServe deployments may continue to receive bug fixes, but the lack of active maintenance, unclear licensing, and explicit deprecation guidance make this unsuitable for new development.

Install

langserve on PyPI

pip install langserve

uv add langserve

poetry add langserve

Installing langserve

Before you install

Low install friction with a pure-Python wheel and four stable runtime dependencies. However, the repository is archived and maintenance is abandoned—bug fixes may be accepted from the community, but no new features will be added.

License in practice

License treatment is unclear; the raw license identifier is 'LangServe' with no SPDX mapping. Verify licensing terms before use in proprietary or restricted contexts.

Quickstart

pip install langserve

from langserve import add_routes
import httpx
import pydantic

add_routes(app, my_runnable)

Requires Python 3.9 or later; integration with a web framework is necessary for server operation.

Verify before relying

• Whether the web framework integration (mentioned in description) is bundled or must be installed separately.
• Current stability and security posture given the archived repository and abandoned maintenance status.
• Whether the LangGraph Platform migration path is straightforward for existing LangServe deployments.

Package facts

Evidence: langserve-0.3.3-py3-none-any.whl

Tags

More WWW/HTTP packages