How to Run an AI Telegram Bot Agent with OE Runtime

Connect your Telegram bot to OE Runtime and automate message monitoring and notifications. The agent reads recent bot updates, summarises incoming messages, and sends AI-composed notifications back to any chat — all from a single YAML file with no code.

Prerequisites: Create a Telegram Bot

1. Open Telegram and search for @BotFather
2. Send /newbot and follow the prompts to name your bot
3. Copy the Bot Token BotFather gives you (format: 123456789:ABCdef...)
4. To find your chat_id: send any message to your bot, then call https://api.telegram.org/bot<TOKEN>/getUpdates in a browser — look for message.chat.id in the response

The Agent YAML

Create a file called agent.yaml inside a telegram/ directory:

name: Telegram Notifier
description: Read recent Telegram messages and send a summary notification
instructions: |
  You are a Telegram bot agent. Read recent updates from the bot and send notification messages.
  Always confirm the bot is connected before reading messages.
  Complete all steps fully before writing your report.
steps:
  - name: Verify Bot Connection
    content: |
      Call the Telegram connector:
      GET /getMe
      Note the bot username, id, and display name to confirm the connection is working.

  - name: Read Recent Updates
    content: |
      Call the Telegram connector:
      GET /getUpdates with params: { "limit": "10", "allowed_updates": "message" }
      For each message note: chat_id, from username, message text, and date.
      Remember the chat_id from the most recent message — you will use it to send the reply.
      If no updates are found, note that the bot has no recent messages and skip the next step.

  - name: Send Summary Notification
    content: |
      Compose a brief summary of the recent messages.
      Use the chat_id from the most recent update (from Step 2).
      Send it via the Telegram connector:
      POST /sendMessage with body:
      {
        "chat_id": <chat_id from Step 2>,
        "text": "<your summary message>",
        "parse_mode": "Markdown"
      }

  - name: Report
    content: |
      Summarize what was done:
      - Bot name and username confirmed
      - Number of recent messages read
      - Summary of message content
      - Notification sent to chat_id (confirm with message_id returned)
connectors:
  - connection_name: My Telegram Bot
    connection_type: telegram

The Config File

Create oe-config.json in the same directory. The Telegram Bot Token is embedded directly in the baseUrl — no separate auth header is required:

{
  "llm": {
    "provider": "openai",
    "model": "gpt-4o",
    "apiKey": "YOUR_OPENAI_API_KEY"
  },
  "server": {
    "enabled": false,
    "port": 3333,
    "apiKey": "your-secret-api-key"
  },
  "connectors": [
    {
      "connection_name": "My Telegram Bot",
      "connection_type": "telegram",
      "baseUrl": "https://api.telegram.org/botYOUR_BOT_TOKEN"
    }
  ]
}

Replace YOUR_BOT_TOKEN with the token from BotFather. The token becomes part of the URL — this is how the Telegram Bot API authenticates requests.

Download OE Runtime

Run the Agent

Windows

# Run the agent (pass your chat_id as a param)
oe-runtime-win.exe telegram/agent.yaml --config telegram/oe-config.json

macOS

# Make executable (first time only)
chmod +x oe-runtime-macos

# Run the agent
oe-runtime-macos telegram/agent.yaml --config telegram/oe-config.json

Linux

# Make executable (first time only)
chmod +x oe-runtime-linux

# Run the agent
oe-runtime-linux telegram/agent.yaml --config telegram/oe-config.json

Running as an API Server

Add --serve to start OE Runtime as an HTTP server on port 3333:

oe-runtime-win.exe --serve --config telegram/oe-config.json

Then trigger the agent via API:

curl -X POST http://localhost:3333/run \
  -H "Content-Type: application/json" \
  -d '{"yaml": "telegram/agent.yaml", "params": {}}'

Use Cases

• Server and infrastructure alerts — Send instant Telegram notifications when a server goes down, a deployment fails, or a health check returns an error. The agent pulls status from your monitoring tools and formats a clear, actionable alert message to your on-call engineer's phone within seconds.
• Daily briefing delivery — Schedule the agent to run each morning and send a structured daily briefing to a Telegram group covering tasks due today, key metrics from the previous day, and any flagged items that need attention. The entire briefing is AI-composed from your connected data sources.
• Customer support intake — Monitor a Telegram support bot for incoming customer messages and automatically route, tag, or respond to them based on content. The agent can handle common questions instantly and escalate complex issues with a structured summary to the right team member.
• Sales pipeline notifications — Notify your sales team on Telegram when a new lead is created in your CRM, a deal moves stages, or a high-value prospect opens a proposal. Keeps the team informed in real time without requiring them to check dashboards.
• CI/CD pipeline status — Post build and deployment status updates to a Telegram channel for your engineering team. When a build passes or fails, the agent sends a formatted message with the branch name, commit author, and a direct link to the pipeline run.
• E-commerce order alerts — Notify fulfilment or warehouse teams via Telegram when high-value orders are placed, stock levels drop below threshold, or a return is initiated. The agent connects to your order management system and composes clear, actionable messages.
• Scheduled report delivery — Run a weekly or monthly report agent that pulls data from your databases or APIs, composes a formatted summary, and sends it directly to a Telegram group or channel. No email required — reports land where the team already communicates.
• IoT and sensor alerts — Forward alerts from connected devices to a Telegram chat when a sensor reading exceeds a threshold, a device goes offline, or an anomaly is detected. The agent formats the raw sensor data into a human-readable message with context and suggested action.