Overview

A Chainlit application can be consumed through multiple platforms. Write your assistant logic once, use everywhere!

​

Available Platforms

​

Tips & Tricks

​

Start Chainlit with -h

When running a Chainlit app in production, you should always add -h to the chainlit run command. Otherwise a browser window will be opened server side and might break your deployment.

​

Double check the host

By default, the Chainlit server host is 127.0.0.1. Typically, if you are running Chainlit on docker, you want to add --host 0.0.0.0 to your chainlit command.

​

Account for websockets

Chainlit is built upon websockets, which means the service you deploy your app to has to support them. When auto scaling, make sure to enable sticky sessions (or session affinity).

Even with sticky sessions, load balancers sometime struggle to consistently route a client to the same container. In that case you can set transports = ["websocket"] in your .chainlit/config.toml file.

​

Deploying Chainlit on a subpath

If you need to deploy your Chainlit app to a subpath like https://my-app.com/chainlit, you will need to set the --root-path /chainlit flag when running the chainlit run command. This will ensure that the app is served from the correct path.

​

Cross origins

If your end users consumes the Chainlit UI from the same origin as the server, everything will work out of the box. However, if you embed Chainlit on a website, the connection will fail because of CORS.

In that case, you will have to update the allow_origins field of your .chainlit/config.toml.

​

Community ressource

After you’ve successfully set up and tested your Chainlit application locally, the next step is to make it accessible to a wider audience by deploying it to a hosting service. This guide provides various options for self-hosting your Chainlit app.

• on Ploomber Cloud
• on AWS
• on Azure Container
• on Google Cloud Run
• on Google App Engine
• on Replit
• on Render
• on Fly.io
• on HuggingFace Spaces