SMS Character Counter

How SMS character counting works

An SMS is not measured in words or in bytes of a file — it is measured in a fixed 140-byte payload defined by the 3GPP standards every mobile network follows, Nepal Telecom and Ncell included. How many characters fit into those 140 bytes depends entirely on which alphabet your text can be encoded in.

GSM-7: the 160-character alphabet

If every character of your message appears in the GSM 03.38 default alphabet — the plain English letters, digits, common punctuation and a handful of European letters — each character is packed into 7 bits instead of 8. Seven bits × 160 characters = 140 bytes exactly. That is where the famous 160-character SMS comes from.

Unicode (UCS-2): the 70-character alphabet

The moment a single character in your message is not in that alphabet — any Nepali letter, an emoji, a curly apostrophe — the whole message must be sent in UCS-2, a 16-bit encoding. Two bytes per character means only 70 characters fit into the same 140 bytes. This is all-or-nothing: one Devanagari character in an otherwise English message drops the limit for the entire message from 160 to 70.

Longer messages: 153 and 67

A message longer than one part is not rejected — it is chopped into several SMS, each carrying a small User Data Header (UDH) that tells the receiving phone how to stitch them back together. That header eats 6 bytes of the 140, so every part of a multipart message holds 153 GSM-7 characters or 67 Unicode characters instead of 160 or 70. The recipient normally sees one long message; you are billed for each part.

Counting Nepali (Devanagari) SMS

This is the part that surprises almost every sender in Nepal, so it is worth being precise. Nepali written in Devanagari — नेपाली — has no place in the GSM-7 alphabet, which contains only Latin letters, Greek capitals and a few symbols. Every Nepali SMS is therefore sent as Unicode, and the working limits are:

• 70 characters in a single Nepali SMS,
• 67 characters per part once the message needs more than one part,
• 1 credit per part, per recipient — a 200-character Nepali notice sent to 1,000 people costs 3,000 credits, not 1,000.

Every matra, halanta and anusvara is its own character

A Nepali reader sees नमस्ते as a few letters. The network sees six separate Unicode characters, because the vowel signs (ा ि ी े ो ौ), the halanta or virama (्) that joins consonants into a conjunct, and the anusvara and chandrabindu (ं ँ) are all independent code points that simply render attached to the letter beside them. Each one uses two bytes on the wire and each one counts against your 70.

Marks are shown on a dotted circle (◌) because they cannot stand alone. Hover any cell to see its Unicode code point.

Nepali digits and the danda cost you the whole message

Three habits quietly force a message into Unicode even when the rest of it is English:

• Devanagari digits ० १ २ ३ ४ ५ ६ ७ ८ ९ are separate characters from 0 1 2 3 4 5 6 7 8 9. The ASCII digits are GSM-7; the Devanagari ones are not.
• The danda । and double danda ॥ are Devanagari punctuation, not a full stop. A single danda at the end of an English sentence makes the entire message Unicode.
• रु and ₹ are Unicode; Rs and Rs. are GSM-7. (The euro sign € is in GSM-7, but costs two characters.)

Romanised Nepali fits more than twice as much

The same sentence written in Devanagari and in Roman letters costs very differently, because the Roman version stays in GSM-7:

Both fit in one SMS here — but the Devanagari version has only 26 characters of room left, while the romanised one has 111. For long notices, receipts and reminders, writing in romanised Nepali routinely turns a 3-credit message into a 1-credit message. Whether that is worth it depends on your audience: Devanagari reads far better for public notices and older recipients, and every handset sold in Nepal renders it correctly.

NTC and Ncell both deliver Unicode SMS

Nepal Telecom and Ncell both accept and deliver UCS-2 messages, and both support concatenated (multipart) SMS, so a long Devanagari message arrives as one continuous message on any modern handset. Two practical notes for Nepal:

• Your sender ID — the alphanumeric name shown instead of a number, up to 11 characters — is carried in a separate field and does not count towards the 160 or 70 characters of the message body.
• A very old feature phone may show a long message as separate parts. It is delivered and charged identically either way.

Worked examples

These are counted by the same code that prices a real send on Kusha SMS:

Characters that cost more than they look

Most “why is my message suddenly two SMS?” tickets come down to one of these. Word processors, phone keyboards and copy-pasted text introduce them without asking.

The full GSM-7 alphabet

If your message uses only these characters, it sends as GSM-7 at 160 characters per SMS. Anything else — including every Devanagari character — makes it Unicode at 70.

Plus the line feed and carriage return used for line breaks. These characters — ^{}\[~]|€ — are also GSM-7 but count as two characters each.

Frequently asked questions

About this counter

This tool is published by Kusha SMS, a bulk SMS platform for Nepal built by Sahasra Tech. It sends campaigns, alerts and OTPs to NTC and Ncell numbers with contacts, templates, scheduling, delivery reports and a REST API. The counter on this page runs the same encoding and segmentation rules the platform uses to price and bill every message, so what you see here is what a send actually costs.

Nothing you type is transmitted: the count is computed in your browser and the page makes no network request while you type.