alibaba/options

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

Alibaba provider configuration options

The Alibaba provider supports the following configuration options: baseURL (string, default 'https://dashscope-intl.aliyuncs.com/compatible-mode/v1'), videoBaseURL (string, default 'https://dashscope-intl.aliyuncs.com'), embeddingBaseURL (string, default 'https://dashscope-intl.aliyuncs.com/api/v1'), apiKey (string, defaults to ALIBABA_API_KEY environment variable), headers (Record<string,string> for custom headers), fetch (custom fetch implementation), includeUsage (boolean, defaults to true for including usage in streaming responses).

Alibaba language model provider options

Alibaba language models support the following provider options: enableThinking (boolean, defaults to false, enables thinking/reasoning mode), thinkingBudget (number, maximum reasoning tokens to generate), parallelToolCalls (boolean, defaults to true, enables parallel function calling).

Alibaba embedding model provider options

Alibaba embedding models support the following provider options: textType ('query' | 'document', defaults to 'document', differentiates query text from document text), dimension (number, defaults to 1024, 'text-embedding-v4' also supports 1536 and 2048), outputType ('dense' | 'sparse' | 'dense&sparse', defaults to 'dense', sparse-only output is not supported and will throw, use 'dense&sparse' to receive dense embeddings in embedding/embeddings and sparse vectors in providerMetadata.alibaba.sparseEmbeddings).

Alibaba video provider options

Video provider options available via providerOptions.alibaba: negativePrompt (string, max 500 characters, description of what to avoid), audioUrl (string, URL to audio file for audio-video sync, WAV/MP3, 3-30 seconds, max 15MB), promptExtend (boolean, defaults to true, enable prompt extension/rewriting), shotType ('single' | 'multi', wan2.6 models only, multi enables multi-shot cinematic narrative), watermark (boolean, defaults to false, whether to add watermark), audio (boolean, for wan2.6 I2V and R2V models; wan2.7 models always generate audio), referenceUrls (string[], array of reference image/video URLs for wan2.6 models, supports 0-5 images and 0-3 videos, max 5 total), media (Array with type, url, and optional referenceVoice for wan2.7 models, overrides automatic mapping from inputReferences and frameImages), ratio ('16:9' | '9:16' | '1:1' | '4:3' | '3:4', wan2.7 text-to-video and reference-to-video models, top-level aspectRatio option is mapped automatically), pollIntervalMs (number, defaults to 5000, polling interval for checking task status), pollTimeoutMs (number, defaults to 600000, maximum wait time for video generation).

Give your agent this brain