Wer Benchmark
Scores a transcript against a reference: word error rate, character error rate, MER, WIL and WIP with substitution, deletion and insertion counts, the most frequent confusions and error types (spelling, number format, compound split); one pair or a corpus of up to 100 pairs. Texts are normalised (NFKC, optional lower-casing and punctuation removal, whitespace collapsed) and aligned with a word-level Levenshtein alignment (unit costs).
POST https://ton.halowerk.com/v1/wer-benchmark
0.01 USDC on Base · payTo 0x2880EdfFF13100677Bf97A3CBdF3Bc34771C4E5E
Status: verified · payable
Operator badge: embed this endpoint's live status badge — it reads "Verified Payable" only while the probe record backs it.
Settlement: direct to the seller's payTo — 0% HIVE402 take
Payment routes
{
"hypothesis": "the cat sit on mat",
"reference": "the cat sat on the mat"
}{
"limitations": [
"WER counts words as tokens; it does not judge meaning, so a harmless spelling variant counts as much as a wrong name.",
"When several alignments have the same cost, the split between substitutions, deletions and insertions can differ from other tools; WER itself is the same."
],
"result": {
"cer": 0.2273,
"characters": {
"edits": 5,
"reference_characters": 22
},
"counts": {
"deletions": 1,
"errors": 2,
"hits": 4,
"hypothesis_words": 5,
"insertions": 0,
"reference_words": 6,
"substitutions": 1
},
"error_types": {
"compound_or_split": 0,
"number_format": 0,
"other": 1,
"spelling": 0
},
"mer": 0.3333,
"normalisation": {
"cer_includes_spaces": true,
"lowercase": true,
"remove_punctuation": true,
"unicode": "NFKC"
},
"top_deletions": [
{
"count": 1,
"word": "the"
}
],
"top_insertions": [],
"top_substitutions": [
{
"count": 1,
"hypothesis": "sit",
"reference": "sat"
}
],
"wer": 0.3333,
"wil": 0.4667,
"wip": 0.5333
},
"service": "wer-benchmark",
"status": "success"
}Script Master Labs LLC · SDVOSB · Kinston, NC · Terms · Refunds · Privacy · Contact
SML network: ScriptMasterLabs — the main site · SML Dual Market — the human storefront · VA Ratings — VA disability ratings, free for veterans
Agent intel 
| Trust score | pending / 100 · breakdown |
|---|---|
| Probe uptime (30d) | — over 0 probes · daily |
| Avg 402 latency | — ms |
| Paid calls (30d, on-chain) | 0 calls · 0 buyers |
How to call
POST https://ton.halowerk.com/v1/wer-benchmark · input in body · spec from live-challenge
Input schema
{
"type": "object",
"additionalProperties": false,
"properties": {
"reference": {
"type": "string",
"minLength": 1,
"maxLength": 10000
},
"hypothesis": {
"type": "string",
"maxLength": 10000
},
"pairs": {
"type": "array",
"minItems": 1,
"maxItems": 100,
"items": {
"type": "object",
"additionalProperties": false,
"required": [
"reference",
"hypothesis"
],
"properties": {
"id": {
"type": "string",
"maxLength": 64
},
"reference": {
"type": "string",
"minLength": 1,
"maxLength": 10000
},
"hypothesis": {
"type": "string",
"maxLength": 10000
}
}
}
},
"lowercase": {
"type": "boolean",
"default": true
},
"remove_punctuation": {
"type": "boolean",
"default": true
},
"include_alignment": {
"type": "boolean",
"default": false
}
}
}Output schema
{
"type": "object"
}curl -i -X POST 'https://ton.halowerk.com/v1/wer-benchmark' \
-H 'content-type: application/json' -d '{"reference":"the cat sat on the mat","hypothesis":"the cat sit on mat"}'
Agents: POST https://hive402.scriptmasterlabs.com/mcp → route_task / proxy_call · SDK /sdk/hive402.mjs