new·The score now tells you which way it movedA brain's exam only ever grows: its own material writes questions, and so does every question a real caller asked and did not get answered. The score is a percentage over that growing set, so a brain that learned more could post a smaller number — and this week three did. One of them answered two MORE questions than the week before and showed eighteen points less. Printed as a single percentage, that reads as decline to a reader and as punishment to anyone who contributes material.all news →
mozg.beta
Sign in

AI SDK · Providers · all subjects

cohere/options

6 notes, read out of this brain and free to use. Each one was extracted from a source and is re-checked against its exam.

Cohere provider customization options

The createCohere function accepts the following optional settings: baseURL (string, default https://api.cohere.com/v2) for API calls proxy, apiKey (string, default COHERE_API_KEY environment variable) for authorization header, headers (Record<string,string>) for custom headers, fetch ((input: RequestInfo, init?: RequestInit) => Promise<Response>) for custom fetch implementation, generateId (() => string) for unique request ID generation.

Cohere image detail provider option

Cohere supports image detail levels via providerOptions.cohere.detail on file parts. Valid values are 'high', 'low', or 'auto'. Example: { type: 'file', mediaType: 'image', data: 'url', providerOptions: { cohere: { detail: 'high' } } }

Cohere embedding model provider options

Cohere embedding models support the following provider options via providerOptions.cohere: inputType ('search_document' | 'search_query' | 'classification' | 'clustering', default 'search_query') specifies the type of input, truncate ('NONE' | 'START' | 'END', default 'END') specifies how to handle inputs longer than maximum token length.

Cohere embedding inputType option values

Cohere embedding inputType option values: search_document (for embeddings stored in vector database for search), search_query (for search query embeddings against vector DB), classification (for embeddings passed through text classifier), clustering (for embeddings run through clustering algorithm).

Cohere embedding truncate option values

Cohere embedding truncate option values: NONE (returns error if input exceeds maximum token length), START (discards start of input until remaining is exactly maximum length), END (discards end of input until remaining is exactly maximum length).

Cohere reranking model provider options

Cohere reranking models support the following provider options via providerOptions.cohere: maxTokensPerDoc (number, default 4096) for maximum tokens per document, priority (number, default 0) for request priority.

Give your agent this brain