Google Chat Samples

This repository provides various implementation patterns for Google Chat apps. Each sample folder contains language-specific implementations (Node.js, Python, Java, Apps Script). Key sample types include:

• Basic app: An HTTP endpoint app (using Google App Engine) that receives event notices/messages and responds synchronously with text.
• Avatar app: An HTTP endpoint app (using Google Cloud Run Functions) that responds synchronously with card-formatted responses.
• Pub/Sub app: An app that uses Google Cloud Pub/Sub to receive messages and responds to Google Chat asynchronously.
• Vote app: Demonstrates updating interactive cards via user voting (text or image-based).
• Contact form app: Demonstrates handling input forms, dialogs, and action parameters using cards.
• Preview link app: Shows how to use preview links to provide metadata for links in messages.
• App home app: Handles app home event notices and responds synchronously with cards (includes initialization and updates).
• Selection input app: Demonstrates using external data sources to dynamically populate selection items in card widgets.
• User Auth app: Demonstrates obtaining authorization to call the Chat API with user credentials and storing tokens in a database.
• Webhook app: Demonstrates sending messages to Google Chat using incoming webhooks.

The AI Knowledge Assistant is a Google Chat app that uses generative AI (Vertex AI with Gemini) to answer questions based on conversation history within Chat spaces.

Key Features:

• Real-time Knowledge Retrieval: Uses the Google Workspace Events API and Pub/Sub to monitor conversations in real time, allowing it to answer questions even when not explicitly @mentioned.
• Dynamic Knowledge Base: It treats all messages in a space as a data source. It searches for existing answers first; if none are found, it @mentions a space manager to request an answer.
• Continuous Learning: The app adapts as it continuously processes new conversations in the spaces where it is added.

Architecture:

• Runtime: Node.js.
• Compute: Implemented as two separate Google Cloud Functions.
• Event Handling: Responds to interaction events and subscription events.

The Google Chat Incident Response app is an Apps Script-based sample that demonstrates how to automate technical incident workflows. It consists of two primary components:

1. Web Page: A simulation interface where users enter incident details (title, description, and responder email addresses).
2. Chat App: An automated agent that receives data from the web page to:
   • Create a new Chat space.
   • Add specified responders as members to the space.
   • Post the incident description as the initial message.

Post-Mortem Automation: Once an incident is resolved, the app supports a slash command that automatically generates a Google Doc post-mortem. This document includes:

• A user-provided description of the resolution.
• A full transcript of the Chat conversation.
• An automated summary of the conversation generated via Vertex AI.