2026-01-12 — Prompt Evaluations v2
Added weighted scorecards, benchmark sets, and evaluation summaries.
Everything needed to authenticate, run prompts, connect tools, and operationalize output quality at scale.
Create an API key in workspace settings, then call the prompt execution endpoint with a template ID and variables payload.
curl -X POST https://api.stylevant.com/v1/prompts/run \
-H "Authorization: Bearer $STYLEVANT_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"template_id": "launch-email-v2",
"variables": {
"audience": "mid-market SaaS buyers",
"value_prop": "faster campaign production"
}
}'
Responses include generated content, evaluation metadata, moderation flags, and usage metrics for reporting.
Added weighted scorecards, benchmark sets, and evaluation summaries.
Expanded admin controls for role assignment and invite restrictions.
Added endpoints to install and manage marketplace packs programmatically.
Create a replacement key, deploy config updates, then revoke the old key after traffic shifts.
Email support@stylevant.com with workspace ID, timestamp, and request ID if available.