Score content

Score content against the model’s traits. URLs in content are auto-fetched and their extracted text is scored. Plain-text bodies (curl convenience: pipe a string or URL with no JSON wrapper) are accepted alongside JSON. Set Accept: text/plain for the bar-chart formatter.

POST /models/{handle}/score

cURL

cURL

curl https://u22a8.ai/v1/models/u22a8.commit-message/score \
  -H "Authorization: Bearer $U22A8_API_KEY" \
  -H "Content-Type: text/plain" \
  --data-binary 'Fix race condition in worker dispatch by draining the in-flight queue before shutdown'

200

{
  "scores": {},
  "composite": 123,
  "confidence": "moderate",
  "headroom": 123,
  "detail": {},
  "input_tokens": 0,
  "metadata": {
    "request_id": "<string>",
    "version_id": "v37",
    "sample_ids": [
      123
    ]
  }
}

Send content as JSON, or the raw text with Content-Type: text/plain. A URL in content is fetched and its extracted text is scored. Set Accept: text/plain for a bar-chart rendering instead of JSON. An optional feedback map captures labels as new samples for continuous learning. The full response shape is documented in the score card.

Authorizations

Authorization · string · header · required

API key issued in the platform console, sent as Authorization: Bearer <key>.

Path Parameters

handle · string · required

Model handle, e.g. 'u22a8.technical-writing'.

Example:

"u22a8.technical-writing"

Body

application/jsontext/plain

content · string · required

Text or URL to score. URLs are auto-fetched.

Example:

"https://example.com/article"

traits · string[] | null

Score only these traits. Omit to score all.

feedback · Feedback · object

Optional {trait_key_or_name: label_or_quality} map for continuous learning. When supplied, each entry is captured as a labelled sample alongside the score. Values may be a polarity label ("positive"|"negative"|"good"|...) or a float in [0, 1]. Rejected with feature_unavailable on models with zero-data-retention enabled.

Show child attributes

Example:

{ "clarity": "positive", "depth": 0.8 }

Response

200

application/json

Successful Response

scores · Scores · object · required

Show child attributes

composite · integer · required

confidence · string · default: moderate

headroom · integer | null

detail · Detail · object

Show child attributes

input_tokens · integer · default: 0

metadata · ScoreMetadata · object

Audit and lifecycle metadata attached to a score response.

version_id is the active snapshot the score served from, formatted as "vN" to match the pinned-handle form (handle@vN); null before the first train. sample_ids is populated only when the request included a feedback block — one id per trait whose feedback was captured.

Show child attributes