Chat SDK adapter for Liveblocks

We’re releasing a new adapter for Chat SDK, enabling you to create cross-platform AI bots that can respond in Liveblocks comment threads.

on April 8th

Product updates Open source AI

Ready to get started?

Join thousands of companies using Liveblocks to build multiplayer experiences for people and agents.

Get started for free

Chat SDK adapter for Liveblocks

We’re releasing a new adapter for Chat SDK, enabling you to create cross-platform AI bots that can respond in Liveblocks comment threads.

on April 8th

Today, we’re releasing a new adapter for Chat SDK that lets you build cross-platform bots that can respond in Liveblocks comment threads. With a single codebase, you can power bots across 10+ platforms, including Slack, Microsoft Teams, Discord—and now Liveblocks.

Trigger AI workflows from other platforms

One common use case for Chat SDK is triggering AI workflows from external apps like Slack. Picture this: you’ve built a realtime app with Liveblocks that lets users create table-based documents. From Slack, you can @mention your bot, and ask it to create a new document.

Triggering an AI change from Slack — Triggering the bot’s “Create document” tool from Slack

With the addition of the Liveblocks adapter for Chat SDK, you can now @mention the same bot inside Liveblocks comment threads, and it behaves identically to Slack—all powered by a single codebase.

Triggering an AI change from Liveblocks comments — Triggering the bot’s “Edit document” tool from Liveblocks

Liveblocks adapter

Chat SDK exposes simple TypeScript events and methods, which you can wire into your adapters. For example, here’s how to create a bot that automatically replies when it’s mentioned.

bot.onNewMention(async (thread, message) => {  await thread.adapter.addReaction(thread.id, message.id, "👀");  await thread.post(`Hello, ${message.author.userName}! Our support t…`);});

Each adapter translates these events into platform-specific actions, so your bot logic stays platform-agnostic. With the Liveblocks adapter configured, this code produces an automated reply inside a comment thread.

You can register multiple adapters simultaneously, allowing a single bot to operate across platforms.

Setting up the adapter

To get started, create an instance of the Liveblocks adapter with createLiveblocksAdapter and pass it into new Chat. You’ll need Liveblocks secret keys and a webhook configured in your dashboard.

import { Chat } from "chat";import { createLiveblocksAdapter } from "@liveblocks/chat-sdk-adapter";import { createMemoryState } from "@chat-adapter/state-memory";
const liveblocks = createLiveblocksAdapter({  apiKey: "sk_prod_xxxxxxxxxxxxxxxxxxxxxxxx",  webhookSecret: "whsec_...",  botUserId: "__bot__",  botUserName: "Liveblocks Bot",});
const bot = new Chat({  userName: "Liveblocks Bot",  adapters: { liveblocks },  state: createMemoryState(),});
bot.onNewMention(async (thread, message) => {  await thread.adapter.addReaction(thread.id, message.id, "👀");  await thread.post(`Hello, ${message.author.userName}! Our support t…`);});

Find more information in our Chat SDK documentation or try our Chat SDK bot example.

Adding AI to your bot

You can add AI to your bot in different ways. One approach is to wait until the bot is mentioned, and then subscribe to the thread and automatically reply to all subsequent messages.

To set up a bot that replies to all comments after its first mention in a thread, use thread.subscribe and bot.onSubscribedMessage.

// First mention in the thread, subscribe to thread and replybot.onNewMention(async (thread, message) => {  await thread.subscribe();  const response = await __generateAIResponse__(message.text);  await thread.post(response);});
// On receiving later thread messages, replybot.onSubscribedMessage(async (thread, message) => {  const response = await __generateAIResponse__(message.text);  await thread.post(response);});

Our Chat SDK AI bot example shows how to set up a bot that replies when mentioned, and describes how to set up a subscription inside bot.ts.

Get started now

To get started, follow our Next.js quickstart and deploy your first cross-platform bot in minutes.

Get started now

Contributors

Contributors include:

1 authors

Product updates Open source AI

Ready to get started?

Join thousands of companies using Liveblocks to build multiplayer experiences for people and agents.

Get started for free

Trigger AI workflows from other platforms

Liveblocks adapter

Chat SDK exposes simple TypeScript events and methods, which you can wire into your adapters. For example, here’s how to create a bot that automatically replies when it’s mentioned.

bot.onNewMention(async (thread, message) => {  await thread.adapter.addReaction(thread.id, message.id, "👀");  await thread.post(`Hello, ${message.author.userName}! Our support t…`);});

You can register multiple adapters simultaneously, allowing a single bot to operate across platforms.

Setting up the adapter

import { Chat } from "chat";import { createLiveblocksAdapter } from "@liveblocks/chat-sdk-adapter";import { createMemoryState } from "@chat-adapter/state-memory";
const liveblocks = createLiveblocksAdapter({  apiKey: "sk_prod_xxxxxxxxxxxxxxxxxxxxxxxx",  webhookSecret: "whsec_...",  botUserId: "__bot__",  botUserName: "Liveblocks Bot",});
const bot = new Chat({  userName: "Liveblocks Bot",  adapters: { liveblocks },  state: createMemoryState(),});
bot.onNewMention(async (thread, message) => {  await thread.adapter.addReaction(thread.id, message.id, "👀");  await thread.post(`Hello, ${message.author.userName}! Our support t…`);});

Find more information in our Chat SDK documentation or try our Chat SDK bot example.

Adding AI to your bot

You can add AI to your bot in different ways. One approach is to wait until the bot is mentioned, and then subscribe to the thread and automatically reply to all subsequent messages.

To set up a bot that replies to all comments after its first mention in a thread, use thread.subscribe and bot.onSubscribedMessage.

// First mention in the thread, subscribe to thread and replybot.onNewMention(async (thread, message) => {  await thread.subscribe();  const response = await __generateAIResponse__(message.text);  await thread.post(response);});
// On receiving later thread messages, replybot.onSubscribedMessage(async (thread, message) => {  const response = await __generateAIResponse__(message.text);  await thread.post(response);});

Our Chat SDK AI bot example shows how to set up a bot that replies when mentioned, and describes how to set up a subscription inside bot.ts.

Get started now

To get started, follow our Next.js quickstart and deploy your first cross-platform bot in minutes.

Get started now

Contributors

Contributors include:

1 authors

Chat SDK adapter for Liveblocks

Ready to get started?

Related blog posts

Multiplayer SDK for React Flow: Realtime collaboration between humans and agents

AI agents are becoming native users of software

Introducing Feeds and APIs for Agent Workflows

Chat SDK adapter for Liveblocks

Trigger AI workflows from other platforms

Liveblocks adapter

Setting up the adapter

Adding AI to your bot

Get started now

Contributors

Ready to get started?

Related blog posts

Multiplayer SDK for React Flow: Realtime collaboration between humans and agents

AI agents are becoming native users of software

Introducing Feeds and APIs for Agent Workflows

Trigger AI workflows from other platforms

Liveblocks adapter

Setting up the adapter

Adding AI to your bot

Get started now

Contributors