Can a serverless Postgres database really handle the demands of a real-time application? The answer lies in pairing it with the right publish-subscribe model. In this guide, you will learn how to combine the real-time capabilities of Ably LiveSync with the structured power of Neon Postgres to build a optimistic and scalable comment system in your Next.js application.
Prerequisites
To follow this guide, you’ll need the following:
- Node.js 18 or later
- A Neon account
- An Ably account
- A Vercel account
Create the Next.js application
Let’s get started by cloning the Next.js project with the following command:
Once that is done, move into the project directory and install the necessary dependencies with the following command:
The libraries installed include:
@ably-labs/models
: A library for working with data models and real-time updates in Ably.@neondatabase/serverless
: A serverless Postgres client designed for Neon.@prisma/adapter-neon
: A Prisma adapter for connecting with Neon serverless Postgres.@prisma/client
: Prisma’s auto-generated client for interacting with your database.ably
: A real-time messaging and data synchronization library.ws
: A WebSocket library for Node.js.
The development-specific libraries include:
prisma
: A toolkit for Prisma schema management, migrations, and generating clients.tsx
: A fast TypeScript runtime for development and rebuilding.
Once that's done, copy the .env.example
to .env
via the following command:
Provision a Serverless Postgres
To set up a serverless Postgres, go to the Neon console and create a new project. Once your project is created, you will receive a connection string that you can use to connect to your Neon database. The connection string will look like this:
Replace <user>
, <password>
, <endpoint_hostname>
, <port>
, and <dbname>
with your specific details.
Use this connection string as an environment variable designated as DATABASE_URL
in the .env
file.
Set up Database Schema
In the file named schema.tsx
, you would see the following code:
The code above defines a function that connects to a Neon serverless Postgres database using a DATABASE_URL
environment variable and sets up the necessary schema for the real-time application. It creates two tables, nodes
and outbox
, to store data and manage message processing, respectively. A trigger function, outbox_notify
, is implemented to send notifications using pg_notify
whenever new rows are inserted into the outbox
table. This ensures the database is ready for real-time updates and WebSocket-based communication.
To run the schema against your Neon Postgres, execute the following command:
If it runs succesfully, you should see Database schema set up succesfully.
in the terminal.
Create the UI for Starting Conversations and Synchronizing Chat History
Create a file named page.tsx
in the app/c/[slug]
directory with the following code:
The code above does the following:
- Defines a
loadConversation
function which calls the/api/c
route to fetch the conversation history based on the particular slug (i.e. the conversation ID). - Uses the
useConversation
hook by ElevenLabs to display the toast when the instance is connected, and to sync the real-time message to Postgres using theonMessage
callback. - Defines a
connectConversation
function that instantiates a private conversation with the agent after obtaining a signed URL using the/api/i
route. - Defines a
disconnectConversation
function that disconnects the ongoing conversation with the agent. - Creates a
useEffect
handler which on unmount, ends the ongoing conversation with the agent.
Next, import the TextAnimation
component which displays different state of the conversation, whether AI is listening or speaking (and what if so).
Finally, add a Show Transcript
button that displays the conversation history stored in Neon to the user.
Now, let's move on to deploying the application to Vercel.
Deploy to Vercel
The repository is now ready to deploy to Vercel. Use the following steps to deploy:
- Start by creating a GitHub repository containing your app's code.
- Then, navigate to the Vercel Dashboard and create a New Project.
- Link the new project to the GitHub repository you've just created.
- In Settings, update the Environment Variables to match those in your local
.env
file. - Deploy.
Summary
TODO - In this guide, you learned how to build a real-time AI voice assistant using ElevenLabs and Next.js, integrating it with a Postgres database to store and retrieve conversation histories. You explored the process of setting up a serverless database, creating a customizable AI agent, and implementing a user-friendly interface with animations and message handling. By the end, you gained hands-on experience connecting various technologies to create a fully functional AI voice assistant application.
Need help?
Join our Discord Server to ask questions or see what others are doing with Neon. Users on paid plans can open a support ticket from the console. For more details, see Getting Support.