Enrichment Forge
Mines any uploaded dataset for candidate KPIs, computes them, feeds them back as enrichment adaptors into a decomposed objective, then runs an RLM pass over the cited context to refine each step into a multiscale agentic trace.
Dataset & objective
Parsed in your browser — rows stay a session variable and never enter a prompt.
Upload a dataset and state the objective — the forge finds the KPIs worth mining and turns them into a refined reasoning trace.