Use “Cohere rerank review brief” in Recued. It provides a read-only Cohere reranking workflow.

How it works 7 steps

Inspect the data fetches, transforms, gates, and output this recipe runs.

Process (7 steps)

■ normalized_query trim

Trim whitespace from setting query

◆ rerank recued-core.cohere.rerank.create

Create rerank on cohere

■ results default

Apply default

■ result_count count

Count items in results

✦ summarize core.ai.summarize

AI generates a concise summary

■ result_table to_table

Format results as a data table

■ card to_summary

Format results as a summary card

▼

Output

summary card

table result_table

Settings 7 configurable

Configurable at install. Defaults shown — change them anytime in Recued.

focus setting = Summarize the strongest and weakest ranked documents for this query. Use only the Cohere rerank scores and indexes, and call out any retrieval quality concerns or next-step query/document refinements.

query setting =

top n setting = 10

documents setting = [object Object]

max length setting = 800

rerank model setting = rerank-v3.5

max tokens per doc setting = 4096