A minimal example of a Telegram Bot running on a Cloudflare Worker.

1. Get your new bot token from @BotFather: https://core.telegram.org/bots#6-botfather
2. Sign up to Cloudflare Workers: https://workers.cloudflare.com/
3. In the Cloudflare Dashboard go to "Workers", then click "Create application" and then "Create worker"
4. Choose a name and click "Deploy" to create the worker
5. Click on "Configure worker" -> "Settings" -> "Variables"
6. Add a new variable with the name ENV_BOT_TOKEN and the value of your bot token from @BotFather
7. Add a new variable with the name ENV_BOT_SECRET and set the value to a random secret. See https://core.telegram.org/bots/api#setwebhook
8. Click on "Quick Edit" to change the source code of your new worker
9. Copy and paste the code from bot.js into the editor
10. Optional: Change the WEBHOOK variable to a different path. See https://core.telegram.org/bots/api#setwebhook
11. Click on "Save and Deploy"
12. In the middle panel append /registerWebhook to the url. For example: https://my-worker-123.username.workers.dev/registerWebhook
13. Click "Send". In the right panel should appear Ok. If 401 Unauthorized appears, you may have used a wrong bot token.
14. That's it, now you can send a text message to your Telegram bot

The bot will send the original message back with Echo: prepended. If you want to change it, look at the function onMessage(). It receives a Message object and sends a text back:

/**
 * Handle incoming Message
 * https://core.telegram.org/bots/api#message
 */
function onMessage (message) {
  return sendPlainText(message.chat.id, 'Echo:\n' + message.text)
}

The file bot2.js contains an improved bot, that demonstrates how to react to commands, send and receive inline buttons, and create MarkdownV2-formatted text.

The file bot3.js contains an improved version that replies inline queries with voice messages. The voice messages should be stored in OPUS format and .ogg in the cloud you most like. The audio files are listed in a JSON array with the following structure in a KV namespace called NAMESPACE and with following content under the key input_files.

Go to Workers & Pages -> KV and create a new namespace. Add a new key input_files and store the JSON structure from below with your own audio files.

Now in Overview -> your-worker -> Settings -> Variables -> KV Namespace Bindings bind the namespace to a variable called NAMESPACE.

 [
    [
      "File Name",
      "URL",
      duration,
      "<tg-spoiler> caption </tg-spoiler>"
    ],
    [
      "test",
      "https://example.com/my_file.ogg",
      5,
      "<tg-spoiler>Description in a spoiler</tg-spoiler>"
    ],
  ]

The bot4 is a bot that randomly reacts to messages received. It demostrates how to use big reactions when the 🎉 emoji gets chosen.

The source-code is licensed under the CC0-1.0 license. If you need a different license, there are also branches with the MIT and GPL 3 license.