StatusSign in to track progress Rerank Retrieved Evidence for an Agent Answer
Select useful, diverse retrieved evidence for grounded agent answers.
DifficultyMedium
RAG54%30mStatusSign in to track progress Split Documents into Retrieval-Safe Chunks
Create retrieval chunks that preserve answerable context across boundaries.
DifficultyMedium
RAG59%20mStatusSign in to track progress Build Retrieval Context with Citations
Combine retrieved chunks into a context block with source citations.
DifficultyMedium
RAG61%20mStatusSign in to track progress Gate Retrieval Context Before the Agent Answers
Filter weak retrieval results and decide whether evidence is strong enough to answer.
DifficultyEasy
RAG57%20mStatusSign in to track progress Rewrite Queries for Better RecallPro
Expand a user query with retrieval-friendly terms while keeping the original intent.
DifficultyHard
RAG42%30mStatusSign in to track progress RAG Support Agent PipelinePro
Build a RAG support flow: rewrite ambiguous follow-ups, retrieve relevant documents, produce grounded answers with citations, and refuse unsupported claims.
DifficultyHard
RAG28%35mStatusSign in to track progress Merge Hybrid Search ResultsPro
Merge ranked results from lexical and vector search into a single deduplicated list.
DifficultyHard
RAG44%30m