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

voyage-ai/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.

Voyage AI provider configuration options

The Voyage provider instance supports the following optional settings: baseURL (string, default: https://api.voyageai.com/v1), apiKey (string, defaults to VOYAGE_API_KEY environment variable), headers (Record<string,string> for custom headers), and fetch (custom fetch implementation, defaults to global fetch function).

Voyage AI inputType option behavior

When inputType is 'null', the embedding model directly converts inputs into numerical vectors. When 'query', the input is treated as a search query with prepended context. When 'document', the input is treated as a document to be stored in a vector database.

Voyage AI outputDtype option support by model

'int8' and 'uint8' are supported by voyage-3-large, voyage-3.5, voyage-3.5-lite, and voyage-code-3. 'binary' and 'ubinary' are supported by voyage-3-large, voyage-3.5, voyage-3.5-lite, and voyage-code-3. 'float' is supported by all models.

Voyage AI multimodal embedding provider options

Multimodal embedding models support the following provider options passed via `providerOptions.voyage`: inputType ('query' | 'document', default: 'query'), outputEncoding ('base64' or null for float32, default: null), and truncation (boolean, default: true).

Voyage AI multimodal inputType behavior

When inputType is 'query', the model prepends 'Represent the query for retrieving supporting documents: ' to inputs. When 'document', it prepends 'Represent the document for retrieval: ' to inputs.

Voyage AI multimodal outputEncoding option

When outputEncoding is null (default), embeddings are returned as a list of floating-point numbers (float32). When outputEncoding is 'base64', embeddings are returned as Base64-encoded NumPy array of single-precision floats.

Give your agent this brain