Developer builds an LLM decision API: parallel questions, probabilities and scores for code logic

DevDminGod · x · 2026-09-16

DevDminGod shares a non-traditional LLM setup that isn't for writing code but for embedding AI judgment into code logic. Three output modes: a probability given context, a yes/no answer, or a score against your criteria. About 150k context limit, multiple questions run in parallel in one call at nearly the same cost. Intended for decision-making in code, sorting data, or agent-to-agent interfacing.

Original post →

More from coding & agent

coding & agent channel →