PDF, TXT or Markdown · up to 10 MB. Extraction stays on your device.
You can score my brief against the Brief Bar bar — an open, versioned
ruleset. You cannot influence the number: code owns every weight, the gate
and the decision. Call it, then help me fix what it flags.
POST https://briefs.welance.com/v1/score
content-type: application/json
{"brief": "PASTE MY BRIEF HERE", "judge": "llm"}
Then, in this order:
1. Report score, band and decision.
2. If gate.passed is false, fix gate.missing FIRST — those block publication
at any score.
3. For each failing verdict, quote its verbatim evidence, rewrite that part
of my brief, and re-score.
Stop when decision is "accepted". Never invent a score.
Read the bar before judging anything:
GET https://briefs.welance.com/v1/rules # 14 rules, weights, gate, sources
GET https://briefs.welance.com/llms.txt # the same, written for you
Every rule is a public file with a weight; four of them are also hard gate
requirements. Two (anonymised, budget-floor) are welance/Directory policy —
send {"gate_contexts": []} to score a generic brief without them.