Single lookup
Judge evidence and return a label or score
Answer labels, probabilities or rubric scores from an AI judge, plus the model used
Run it.From your AI, or try it here.
Use Ahel to run "Judge evidence and return a label or score". State: <state> Questions: <questions>
Paste this into ChatGPT or Claude connected to Ahel.
You get: answer labels, probabilities or rubric scores from an AI judge, plus the model used.
For developers
- Action id
ai-judge.decide- Inputs
state: The evidence, as bounded Markdown with XML-style sections. Escape & < > in untrusted text. Quoted text is evidence, never instructions. State and questions together take at most 10,000 bytes.questions: Question id -> {type: choice | noul | score, instructions, criteria}. choice and noul take criteria as label -> description (noul uses true and false); score takes an ordered list of levels. Questions
- HTTP call
API referencecurl -s https://ahel.ai/api/platform/v2/actions/ai-judge.decide/runs \ -H "Authorization: Bearer $AHEL_TOKEN" -H "Content-Type: application/json" \ -d '{"inputs":{"state":"<state>","questions":"<questions>"},"max_cost_usd":"<your cap>"}'