Text splitter

The SMS limit is the one that catches people, because it is about encoding rather than about characters. A single message holds 160 characters while the text stays inside the GSM alphabet; one character outside it. An emoji, a curly apostrophe, an em dash; switches the whole message to 16-bit encoding and the limit drops to 70. A message that fitted comfortably becomes three, at three times the cost. Running the text through the cleaner first, which flattens curly quotes and dashes back to their ASCII equivalents, often brings it back under one segment. A message long enough to be split carries a joining header in every part, so each part holds slightly less than a single message would.

The splitting itself measures the way a browser measures a string, so an emoji counts as two rather than one. Platform counters do not always agree: X counts a link as a fixed length however long it really is, and weights some scripts differently, so treat a chunk size as a ceiling with a little headroom rather than an exact match. The break point is a space: a single word longer than the chunk, which in practice means a URL, gets cut mid-word, because the alternative is an oversized chunk that will be rejected. Chunks are trimmed, so a line break that lands on a boundary is lost.