Attest tools · Text Stats: words, characters, tokens and SMS segments
Count a sample prompt with the bundled o200k_base and cl100k_base tokenizers and check a 50-token budget.
{
"text": "Summarize these inventory changes: cotton sheets increased from 12 to 18 units; linen towels decreased from 9 to 7 units. Return two concise bullet points.",
"limits": [
{
"name": "Prompt budget",
"unit": "tokens_o200k",
"max": 50
}
]
}{
"characters": 155,
"code_points": 155,
"utf16_units": 155,
"bytes_utf8": 155,
"words": 25,
"sentences": 2,
"paragraphs": 1,
"lines": 1,
"whitespace": 24,
"reading_time_minutes": 0.11,
"speaking_time_minutes": 0.14,
"time_basis": {
"reading_words_per_minute": 238,
"speaking_words_per_minute": 183
},
"tokens": {
"o200k_base": 35,
"cl100k_base": 35,
"note": "OpenAI tokenizers, per the model table in js-tiktoken 1.0.21: o200k_base is listed for GPT-4o, GPT-4o mini, GPT-4.1, GPT-4.5 Preview, o1, o3, o4-mini and GPT-5; cl100k_base for GPT-4, GPT-4 Turbo and GPT-3.5 Turbo. Other models, including other vendors', may tokenize differently, so treat these counts as estimates for them."
},
"sms": {
"encoding": "GSM-7",
"segments": 1,
"characters_per_segment": 160,
"units": 155,
"max_segments": 255,
"within_concatenation_limit": true,
"note": "GSM 7-bit default alphabet without national language shift tables; extension characters (^ { } [ ] ~ | \\ € and form feed) cost two. Code 0x09 is capital Ç in the 3GPP table, so small ç is counted as UCS-2, the encoding a gateway following that table sends; a gateway following Unicode's GSM0338.TXT accepts ç as one septet and may need fewer segments than reported. Multipart sizes assume (truncated)"
},
"limits": [
{
"name": "Prompt budget",
"unit": "tokens_o200k",
"max": 50,
"value": 35,
"within": true
}
],
"sources": [
{
"name": "Brysbaert, M. (2019). How many words do we read per minute? A review and meta-analysis of reading rate. Journal of Memory and Language, 109, 104047.",
"url": "https://doi.org/10.1016/j.jml.2019.104047",
"used_for": "reading_time_minutes (238 wpm, adult silent reading of non-fiction) and speaking_time_minutes (183 wpm, adult reading aloud)"
},
{
"name": "OpenAI tiktoken encodings o200k_base and cl100k_base (MIT), rank files from js-tiktoken (MIT)",
"url": "https://github.com/openai/tiktoken",
"used_for": "tokens"
},
{
"name": "3GPP TS 23.038 (GSM 7-bit default alphabet and extension table) and 3GPP TS 23.040 clause 9.2.3.24.1 (concatenated short messages)",
"url": "https://www.3gpp.org/ftp/Specs/archive/23_series/",
"used_for": "sms"
},
{
"name": "Unicode GSM0338.TXT mapping — cited only for the divergence at GSM code 0x09, which it maps to small c-cedilla where the 3GPP table draws the capital; we follow 3GPP",
"url": "https://www.unicode.org/Public/MAPPINGS/ETSI/GSM0338.TXT",
"used_for": "sms"
}
],
"notice": "General information only. Reading and speaking times are averages for adult English readers; individual rates vary.",
"checked_at": "2026-10-20T20:47:48.390Z"
}curl -s -X POST https://getattest.com.au/api/v1/text/stats \
-H "authorization: Bearer $ATTEST_API_KEY" \
-H "content-type: application/json" \
--data-raw '{"text":"Summarize these inventory changes: cotton sheets increased from 12 to 18 units; linen towels decreased from 9 to 7 units. Return two concise bullet points.","limits":[{"name":"Prompt budget","unit":"tokens_o200k","max":50}]}'Run this call with your own bearer key. Issue an API key.
The input was validated and run locally through the registered service. The output above is a display preview of the captured service result. File bytes are described instead of printed, and strings longer than 400 characters are truncated; the HTTP API also adds call_id, credits_charged and credits_remaining. Timestamps and generated identifiers will vary.
Source: apps/api/svc/text_stats.mjs