How Many Characters Can a Telegram Message Hold?

A text message sent by a Telegram bot must be 1 to 4,096 characters after formatting markup is parsed, and the caption attached to a photo or file must be 0 to 1,024 characters. This article uses Korean text as its example, so Korean example sentences are kept as written. One Korean syllable counts as 1 character, so Korean text alone can fill 4,096 characters of message text or 1,024 characters of caption. The values below come from the public Telegram Bot API documentation, and Korean text and emojis were counted with the Geuljasu count functions to find where each limit is hit.

Woman typing a long message on her phone in a cafe AI-generated image

How many characters can a Telegram message hold?

In the Telegram Bot API documentation, the text parameter of the sendMessage method is described as “1-4096 characters after entities parsing”, meaning 1 to 4,096 characters once the formatting has been interpreted. The caption parameter of methods that send a photo (sendPhoto) or a file (sendDocument) is “0-1024 characters after entities parsing”, meaning 0 to 1,024 characters. A caption is optional, so it starts at 0, while a text message cannot be empty, so it starts at 1.

ItemLimit in the official documentationIn Korean characters
Text message (text)1 to 4,096 charactersup to 4,096 Korean characters
Photo or file caption (caption)0 to 1,024 charactersup to 1,024 Korean characters

The text limit is exactly four times the caption limit. When you want to attach a long explanation to a photo, it leaves more room to attach a short caption and send the long explanation as a separate text message.

These values are the documentation’s limits for bots sending through the API. The limit for what a person types in the app’s input box could not be confirmed in the official documentation this time, so do not apply the numbers in this article to the app as they are.

How does that compare with other services?

Discord’s developer documentation gives the message content as “up to 2000 characters” and the embed description as up to 4,096 characters. The Telegram text limit and the Discord embed description limit are both 4,096, and Discord’s regular message content is about half of that.

Character limits by serviceTelegram text: 4,096chars; Discord embed description: 4,096chars; Discord content: 2,000chars; Telegram caption: 1,024chars01,0002,0003,0004,000Telegram text4,096charsDiscord embeddescription4,096charsDiscord content2,000charsTelegram caption1,024chars
Character limits by service — Values from each service's official developer documentation (checked 2026-11-15)

If you post the same text to both services, fit it to the shorter limit. A text longer than 2,000 Korean characters can fit in one Telegram message but not in Discord message content. Counting 2,000 Korean characters with the Geuljasu function gives 2,000 characters and 6,000 UTF-8 bytes.

What does “after entities parsing” mean?

The documentation states the limits as characters “after entities parsing”. When you send formatting such as bold as HTML tags or Markdown symbols, Telegram reads those symbols and turns them into formatting, and the limit appears to be measured on the characters that remain after that interpretation. For example, “안녕” (hello) wrapped in a bold tag is 9 characters as the input string, but after the formatting is interpreted the text is just the 2 characters of “안녕”. With the Geuljasu function, the input string length is 9 and the text after interpretation is 2.

So for text with a lot of formatting, do not compare the length you see in the input box directly with the limit. Conversely, how the address part of a link wrapped in formatting counts toward the characters could not be confirmed this time. Leave some margin near the limit.

How many Korean characters fit?

One Korean syllable is 1 character. Counting a text made of “가” repeated 4,096 times with the Geuljasu function gives 4,096 characters, and one more character makes 4,097, over the limit. For captions the boundary is between 1,024 and 1,025 characters. The documentation states the limit only in characters and gives no byte limit, so the UTF-8 bytes (12,288 bytes for 4,096 Korean characters, 3,072 bytes for 1,024 caption characters) are only reference values and are not used to judge the limit. If you write only Korean, counting characters including spaces and line breaks is enough. You can check the count with spaces at once in the Word Counter.

How do the limits change when emojis are mixed in?

There is a point to check here. The documentation states the unit of 4,096 only as “characters”, while the “Entities” page of the same documentation says to measure the position and length of formatting in “UTF-16 code units”. In UTF-16 code units a Korean character is 1, but an emoji is 2 or more per character. The documentation does not say whether the limit is judged on visible characters (graphemes) or on UTF-16 units, and I could not confirm it by actually sending messages this time. Here are the values counted both ways side by side.

UTF-16 code units in 4,096 charactersEnglish 4,096: 4,096units; Korean 4,096: 4,096units; Smiling-face emoji 4,096: 8,192units; Flag emoji 4,096: 16,384units; Family emoji 4,096: 45,056units010,00020,00030,00040,00050,000English 4,0964,096unitsKorean 4,0964,096unitsSmiling-faceemoji 4,0968,192unitsFlag emoji 4,09616,384unitsFamily emoji4,09645,056units
UTF-16 code units in 4,096 characters — The visible characters are all 4,096, the same (directly calculated values)

For English and Korean the two ways give the same value, so there is nothing to worry about. When emojis are mixed in, the values diverge.

Text compositionCharacters (graphemes)UTF-16 unitsAgainst 4,096
4,096 Korean characters4,0964,096Pass either way (equal to the limit)
4,097 Korean characters4,0974,097Over either way
2,048 smiling-face emojis2,0484,096Pass either way
4,096 smiling-face emojis4,0968,192Pass by characters, over by UTF-16
4,000 Korean characters + 48 smiling-face emojis4,0484,096Pass either way
4,000 Korean characters + 96 smiling-face emojis4,0964,192Pass by characters, over by UTF-16

For a text of Korean mixed with a few smiling-face emojis, the difference between the two ways can grow to about 96 around 4,000 characters. The safest approach is to count a long, emoji-heavy text in UTF-16 units too and keep it at 4,096 or below. A text that passes when counted this way does not exceed the limit however it is judged.

How do I split a text longer than 4,096 characters?

The documented limit is for one text message, so a longer text has to be split into several messages that each stay within the limit. Splitting 10,000 Korean characters into 4,096-character pieces gives three messages of 4,096, 4,096 and 1,808 characters. If you wanted to put 3,000 Korean characters in a caption, it would split into 1,024, 1,024 and 952 characters, so it is better to keep only the opening part in the caption and send the rest as a text message.

Check order before a Telegram bot sends a messagePrepare the text → Photo or file caption?; Photo or file caption? → 1,024 or fewer? (Yes); Photo or file caption? → 4,096 or fewer? (No); 1,024 or fewer? → Send (Yes); 1,024 or fewer? → Split within the limit (No); 4,096 or fewer? → Send (Yes); 4,096 or fewer? → Split within the limit (No); Split within the limit → Photo or file caption?YesNoYesNoYesNoPrepare the textPhoto or filecaption?1,024 or fewer?4,096 or fewer?Split within thelimitSend
Check order before a Telegram bot sends a message — 1,024 characters for a photo caption, 4,096 for a text message

When splitting, it matters not to cut in the middle of a character. Cutting an emoji made of several joined parts in the middle can make it look broken. To preview the part that fits within the limit from the start, enter the limit in the Character Limit Trimmer and paste your text; the point where it goes over is shown right away. If you are curious about bytes, the Byte Counter shows them on a UTF-8 basis.

Code to check it yourself

Paste it into the browser console to see the visible character count and the UTF-16 unit count together, and to split by characters to fit the limit. It uses the same rules that produced the values in the table above.

// Count visible characters (graphemes) and UTF-16 units together
const seg = new Intl.Segmenter('ko', { granularity: 'grapheme' });
function telegramCount(text) {
  return { graphemes: [...seg.segment(text)].length, utf16: text.length };
}

// Split into pieces of `limit` characters (never cut in the middle of a character)
function splitByGraphemes(text, limit = 4096) {
  const parts = [];
  let cur = '', n = 0;
  for (const { segment } of seg.segment(text)) {
    if (n === limit) { parts.push(cur); cur = ''; n = 0; }
    cur += segment; n++;
  }
  if (cur) parts.push(cur);
  return parts;
}

const smile = '\u{1F600}';
telegramCount('가'.repeat(4000) + smile.repeat(96)); // { graphemes: 4096, utf16: 4192 }
splitByGraphemes('가'.repeat(10000)).map((p) => p.length); // [4096, 4096, 1808]

Frequently asked questions

Is a message typed directly in the app also limited to 4,096 characters?

The 4,096 and 1,024 characters in this article are values written in the Telegram Bot API documentation, the limit for messages sent by a bot. The limit for messages a person types in the app’s input box could not be confirmed in the official documentation. If the text will be sent from the app, paste it in and check for yourself.

Do line breaks and spaces count toward the characters?

Counted with the Geuljasu function, one line break is 1 character. A Windows-style line break (CRLF) is 1 character but 2 UTF-16 units. How Telegram counts line breaks toward the limit is not stated in the documentation and could not be confirmed, so leave some margin near the limit.

What if a photo caption is over 1,024 characters?

A caption can hold up to 1,024 characters, so if the explanation is longer, you can write only the key points in the caption and send the rest as a separate text message (up to 4,096 characters). What response Telegram gives when the limit is exceeded could not be confirmed this time. Check the character count first with the Word Counter before sending.

Basis: Telegram Bot API documentation "sendMessage" and "sendPhoto" (text 1-4096 characters and caption 0-1024 characters both measured after entities parsing; latest Bot API listed on the page is 10.4), Telegram developer documentation "Entities" (entity length is in UTF-16 code units), Discord developer documentation "Message Resource" and "Embed Limits" (content 2000 characters and embed description 4096 characters), Geuljasu character and byte count functions. Sources checked on 2026-11-15. Byte values in this post are calculated the same way as the Geuljasu tools; real SMS services may follow different rules set by carriers and sending providers.

The full counting rules are on Methods and standards.

Tools related to this post