Build a Personal Calendar Assistant with Telegram and Agent Capsules

Managing your calendar through multiple apps is tedious. You check Google Calendar on desktop, get notifications on mobile, and manually coordinate across platforms. What if you could schedule meetings, check availability, and get reminders through a single Telegram chat?

In this tutorial, you'll build a personal calendar assistant that handles your Google Calendar through natural conversation on Telegram. The assistant understands requests like "Schedule a team meeting tomorrow at 2pm" or "What's on my calendar this week?" without requiring you to manage servers, databases, or scaling infrastructure.

The calendar assistant you'll build in this tutorial

Agentic Infrastructure

Deploying AI agents to production requires managing extensive infrastructure:

• Provision servers like EC2 instances and configure load balancers

• Deploy and maintain vector databases (Pinecone, Weaviate, or Redis)

• Configure SSL/TLS certificates and API authentication

• Manage secrets securely

• Add monitoring and logging for observability and error tracking

• Configure autoscaling policies and resource limits

Code Capsules eliminates this burden. The platform handles infrastructure, databases, security, and scaling automatically, letting you focus on building Agent tools and features.

Prerequisites

To follow this tutorial, you need:

• A GitHub account and Git installed

• A Code Capsules account

• An LLM API key. For example: Anthropic or OpenAI

Set Up a Redis Capsule

The agents in this tutorial need a vector database. For simplicity, we use a Redis Capsule.

Note: This setup is not recommended for production. Redis is designed for temporary information and caching. For production deployments, consider dedicated vector databases like Pinecone, Weaviate, or Qdrant.

Follow the existing guide to creating a Redis Capsule.

Once you've created the Redis Capsule, copy the connection string from the Capsule Details page.

Configure the Telegram Agent

To set up the Telegram agent, go to the Telegram calendar agent GitHub repository, and click Use this template to create a copy in your own GitHub account.

Then, deploy the Agent Capsule:

1. In your Code Capsules dashboard, click New Capsule and select Agent Capsule.

2. Choose your Team and Space.

3. Connect your GitHub account and select the repository you just created.

4. Click Create Capsule.

For more detailed instructions, see the Agent Capsule deployment guide.

Create a Telegram Bot

To create a Telegram Bot, open the Telegram application and search for BotFather. Start a conversation with BotFather and select /newbot. When prompted, give the bot a name, then a username. Once it is created, the bot provides you with a token. Copy this token and save it for later.

To verify the setup, make sure you can start a conversation with the bot (it won't respond yet).

Now that you have the token, go to your Agent Capsule Config page and add the following variables:

TELEGRAM_BOT_TOKEN=your_bot_token
REDIS_URL=your_copied_connection_string

Paste the Telegram bot token and the Redis connection string as new Environment Variables.

The Telegram agent is now ready. Let's configure the calendar capabilities.

Configure the Google Calendar API

To ensure the agent can access your calendar, you must configure Google Cloud to integrate with it.

Visit the Google Calendar API quickstart. Scroll down and click the Enable the API button.

The button redirects you to the Google Console platform page to confirm your project and enable the API.

After that, navigate to the Configure the OAuth consent screen section in the Google Calendar configuration doc. Click the Go to Branding button, and follow the steps in the Google documentation. If you can't select Internal for the audience, you can select External, but you'll have to add test users with the Google email address you're using.

Then, authorize your web application credentials by clicking the Go to Clients button.

When doing so, set the redirect URI to the URL of the Calendar Agent and specify the endpoint /api/calendar/auth/callback. To find the public URL of your Calendar Agent, open your Agent Capsule dashboard, navigate to the Details tab, and copy the Public URL.

For example, the full URL would have the following structure:

https://agent-capsule-123.ovh-test.ccdns.co/api/calendar/auth/callback

After completing setup, Google provides you with credentials such as the client ID and the client secret. Add these in the Agent Capsule dashboard Config tab as follows:

GOOGLE_CALENDAR_CLIENT_ID=your_client_id
GOOGLE_CALENDAR_CLIENT_SECRET=your_client_secret

Your environment variables configuration should look as follows:

Test the calendar configuration by clicking on the Chat tab, where you can make the following request:

Get the list of events in the calendar for February in the first 3 days.

The Agent should give you a link to click to authorize access to the Calendar API.

After granting access, you should see the following screen:

After the validation, return to the Chat and tell the agent to proceed. You should receive a similar response to this:

Test the Integration

To test the integration, go to Telegram and start a conversation with the bot. Ask the bot to retrieve your calendar events for a particular period. If the bot prompts you for login or authorization, click on the link and grant it access.

After granting access, remind the Agent to continue.

You can also add events to the calendar.

Conclusion

In this tutorial, you built a Telegram bot that acts as an interface to a Code Capsules Agent with Google Calendar capabilities. The Agent uses an LLM to understand your requests and interact with your calendar through natural language.

This same pattern can enable you to build interesting integrations like:

• Task Management Bots: Connect Todoist, Asana, or Jira to manage tasks through chat.

• More Personal Assistants: Combine multiple services like email, calendar, and notes in one conversational interface.

• Customer Support Bots: Integrate with CRM systems and knowledge bases to handle support queries.

• Data Analysis Tools: Query databases or analytics platforms using natural language.

The Agent Capsule architecture makes it easy to add new tools and capabilities without managing infrastructure. You can extend your bot by creating new tools that call external APIs, process data, and integrate with other services.

If you're interested in learning more about building with Agent Capsules, check out the Code Capsules documentation.