Recommend

Run every RAG strategy against the same corpus and questions to find the best-performing configuration.

pdf · docx · pptx · html · csv · json · xlsx · md · txt · images

Or pick a bundled/benchmark dataset

Or paste raw text (one document per line)

Questions

one per line — suffix with ::answer for a reference

0/0

Generator modelEmbedding modelJudge model (LLM-as-judge)

Metric presetQuality weight0.70Cost weight0.15Latency weight0.15