langgraph-api

What it is and what it does

LangGraph API is a local development server that lets you build and test LangGraph agents with a tight feedback loop. It provides an in-memory data store (persisted to local disk on restart) and a browser-based UI for inspecting agent state, runs, and threads. The server supports hot-reload, so code changes trigger automatic restarts without losing your session context.

The package is designed for rapid iteration during development and testing—not production deployment. It depends on a large LLM-agent stack (langchain-core, langchain-protocol, langgraph, starlette, uvicorn, and observability tools like OpenTelemetry). You start it via the langgraph-cli tool with the inmem extra, point it at a langgraph.json config file, and interact with your agent through the UI or programmatically. Debugging support includes remote debugger attachment for IDE breakpoint stepping.

Use it for:

• Iterate on agent logic during development with instant feedback and automatic server reload on code changes.
• Inspect agent state, thread history, and run traces through the browser UI to debug agent behavior.
• Test multi-turn conversations and agent decision-making without setting up a production database.
• Attach an IDE debugger to step through agent code line-by-line during execution.
• Prototype and validate LangGraph workflows before deploying to production infrastructure.

Worth the install?

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

Runs a local development server for LangGraph agents with an in-memory data store, hot-reload support, and a browser-based UI for testing and debugging agent behavior.

Yes, if you are developing LangGraph agents and want a lightweight local testing environment with UI inspection and hot-reload. The low install friction and active maintenance make it a natural fit for the development phase. However, verify that the Elastic License 2.0 terms align with your deployment plans, and note that this is explicitly not for production use—production deployments require separate deployment options.

Install

langgraph-api on PyPI

pip install langgraph-api

uv add langgraph-api

poetry add langgraph-api

Installing langgraph-api

Before you install

Low friction: pure Python wheel with no compiled dependencies. Active maintenance (released 2026-08-13, 1 day old). Requires Python 3.11+. Depends on 34 runtime packages including langchain-core, langgraph, starlette, and uvicorn—a substantial but standard LLM-agent stack.

License in practice

Licensed under Elastic License 2.0 (license_treatment: unclear). Elastic 2.0 is a source-available license with commercial restrictions; verify compatibility with your use case before deploying in production or closed-source contexts.

Quickstart

pip install -U 'langgraph-cli[inmem]'
langgraph dev --config ./langgraph.json
# Server launches with browser UI; state persists in memory during session

Requires Python 3.11 or later. Intended for development/testing only; production deployments require separate deployment options backed by production-grade databases.

Verify before relying

• Whether Elastic License 2.0 restrictions apply to your intended use (commercial, closed-source, or SaaS deployment).
• Performance characteristics and concurrent request limits under typical development workloads.
• Data persistence behavior across server restarts and whether local disk persistence is suitable for your testing workflow.

Package facts

Evidence: langgraph_api-0.12.4-py3-none-any.whl

Tags

More Application Frameworks packages

Further reading