FAQ

Short answers to recurring questions. For the long form, follow the links into the guides.

Does the library escape message text for me?

No — and it does not need to. This library builds keyboards, not message text. MenuBuilder produces an InlineKeyboardMarkup (or, with to_raw(), a plain Bot API dict); the message body you pass to reply_text / edit_message_text is entirely yours. Button labels are sent as literal text and are never parsed as Markdown or HTML, so they need no escaping either.

Escaping only matters for the message text when you send it with a parse_mode, and that is the caller's responsibility. Use the stdlib and python-telegram-bot helpers rather than hand-rolling it:

import html

from telegram.helpers import escape_markdown

# MarkdownV2 — escape every dynamic fragment you interpolate.
safe_md = escape_markdown(user_name, version=2)
await update.message.reply_text(
    f"Welcome, {safe_md}\\!", parse_mode="MarkdownV2"
)

# HTML — stdlib html.escape covers the reserved characters.
safe_html = html.escape(user_name)
await update.message.reply_text(
    f"Welcome, <b>{safe_html}</b>", parse_mode="HTML"
)

Always pass version=2 to escape_markdown

telegram.helpers.escape_markdown defaults to the legacy Markdown rules. Telegram's current dialect is MarkdownV2, which escapes a larger reserved set (_ * [ ] ( ) ~ > # + - = | { } . !), so passversion=2explicitly to match theparse_mode="MarkdownV2"you send. For HTML, the stdlibhtml.escape` is enough.

Plain text needs nothing

If you do not pass a parse_mode, Telegram treats the message as literal text and there is nothing to escape — the simplest and safest default.