Turn your LLM app into an API server instantly with LangServe

Getting to production just got easier. Built with FastAPI, LangServe gives you an API, docs, and a playground for your LangChain apps.

LangServe makes deploying and maintaining your application simple. Get turnkey visibility into usage, errors, performance, and costs when you ship within the LangSmith platform.

Hosted LangServe is Currently in Beta.

Deploy with confidence

Go to Docs

Library of reference examples

Discover, download, and remix templates for a wide range of use-cases ready to be deployed using LangServe.

Quickly iterate 
and collaborate

Hosted LangServe comes with a built in playground and docs site that you can share with your colleagues to try out the application and get feedback faster.

Easy to get started

Hosted LangServe integrates with GitHub, 
so once you connect LangServe to your repository, the service will take care 
of spinning up a container to run your application.

Observability built in

Hosted LangServe is deeply integrated into LangSmith by default. All inputs / outputs from your server are automatically logged to LangSmith, so you can easily debug issues and understand your chain’s behavior.

Native regression testing

Hosted LangServe will show you the exact branch and commit deployed at any given time and how that version of your application is performing.

Auto documentation

Get documentation for your LangServe application built automatically, so that your users and collaborators can take advantage of all the supported endpoints.

The quickest path to production

Go to Docs

One-click deploy

The easiest way to deploy LLM applications. Connect your GitHub repo and deploy in a single click, with LangSmith tracing and observability automatically built in on every request.

Go to Docs

Got a question?

Where is my application hosted?

Hosted LangServe applications are hosted in GCP us-central-1.

What endpoints will my application support out of the box?

You will receive access to invoke, batch, and stream endpoints that mirror the functionality of the Runnables API. You will also receive access to a feedback endpoint where you can add user feedback to runs. For a full list of endpoints, please click on your deployment's URL in your hosted LangServe application, which will redirect to the accompanying docs page.

Do you support an SLA or failover for Hosted LangServe?

Not at the moment since the product is still in beta.

How big are Hosted LangServe instances?

We support up to 1GB of memory on each deployment. If you have a use case that requires more memory, please reach out to hosted-langserve@langchain.dev with your specific use case.

Ready to start shipping 
reliable GenAI apps faster?

LangChain and LangSmith are critical parts of the reference 
architecture to get you from prototype to production.