What Is Inline Mode in Telegram Bots and How to Use It

How inline mode works

Inline mode lets people use a bot without opening a chat with it. In any conversation, group or channel, the user types @botname and a query into the input field, for example @shopbot sneakers. A list of results appears above the keyboard, and the chosen result is sent to the current chat marked "via @botname".

Key properties:

• the bot does not need to be a member of the chat where it is called;
• the bot cannot see the conversation, only the query text and user info;
• the user sends the message, so the other people see it from that user.

How it works under the hood

1. Enable the mode in BotFather with /setinline, where you also set the placeholder hint.
2. As the user types, the bot receives an inline_query update: the query text, an offset for pagination and usually the type of chat the user is typing in.
3. The bot replies with answerInlineQuery, sending up to 50 results per answer.
4. To learn which result was picked, enable /setinlinefeedback; you then receive chosen_inline_result updates.

Useful answer parameters:

• cache_time: how many seconds Telegram caches the results on its side;
• is_personal: cache per user when results are personalized;
• next_offset: load the next batch as the user scrolls;
• button: a button above the results that opens a private chat with the bot or a Mini App, for example to sign in.

Result types

Media also have "cached" variants that use the file_id of a file already on Telegram's servers, which is faster and more reliable than external URLs. Any result can carry an inline keyboard, for example an "Open in store" button.

Business uses

• Sharing products. A shopper searches through @shopbot, sends a card with photo and price to a friend or family chat, and a button leads to the bot or Mini App to order.
• Knowledge base and support. A support agent drops a ready answer or a link to a guide straight into the customer conversation.
• Referral links and promo codes. Users send a personal invitation into any chat.
• Content lookup. Articles, recipes, schedules, exchange rates: anything people discuss in chats.
• Group decisions. Sending a booking slot, a poll or a meeting address into a group chat.

Common mistakes

• Slow answers. Users wait for results on every keystroke. Search an index rather than scanning, and avoid heavy external API calls per character.
• Nothing on an empty query. When the user has typed only @botname, show popular or recent items.
• Wrong caching. Personal prices or account data without is_personal can be served to other users from the cache.
• Treating buttons like normal messages. An inline message has no regular message object; callback queries carry an inline_message_id, and edits must use it.