Version:
interfaceextends Omit<UseVisionProviderOptions, "defaultModel">
UseOllamaVisionProviderOptions
Options for useOllamaVisionProvider.
Accepts all provider config options plus an optional model override.
typescript
interface UseOllamaVisionProviderOptions extends Omit<UseVisionProviderOptions, "defaultModel">Properties
| Name | Type | Description |
|---|---|---|
anthropic? | { apiKey: string; baseURL: string } | Anthropic provider options. |
coordinateScale? | number | Upper bound of the normalized coordinate range (coordinates are in[0, coordinateScale]). Defaults to 1000. Only used whencoordinateSpace is "normalized". |
coordinateSpace? | VercelAiSdkVisionCoordinateSpace | Coordinate space the model reports regions in. Defaults to"pixels". See VercelAiSdkVisionCoordinateSpace. |
defaultConfidence? | number | Default confidence threshold in the range [0, 1]. |
defaultMaxMatches? | number | Default max number of returned matches. |
matching? | VisionMatchingSettings | Explicit matcher tuning settings. When present, these values override equivalent top-level options. |
model? | string | Ollama model ID to use (defaults to llama3.2-vision:11b). |
ollama? | { baseURL: string } | Ollama provider options. |
openai? | { apiKey: string; baseURL: string } | OpenAI provider options. |
openaiCompatible? | { apiKey: string; baseURL: string; headers: Record<string, string>; name: string; queryParams: Record<string, string>; supportsStructuredOutputs: boolean } | Configuration for self-hosted servers that expose the OpenAI Chat Completions API (e.g. vLLM, LMStudio, llama.cpp, LocalAI). Required when the provider is set to "openai-compatible". |
Was this page helpful?