Quiver

Generate and vectorize SVGs

QuiverAI is an AI-powered SVG generation platform that creates high-quality, scalable vector graphics from text descriptions or by vectorizing raster images. It produces clean, resolution-independent SVGs that are ideal for icons, illustrations, logos, and UI elements.

With Quiver, you can:

  • Generate SVGs from text prompts: Describe the vector graphic you need and get production-ready SVG output
  • Vectorize raster images: Convert PNG, JPG, and other raster images into clean SVG vector format
  • Provide reference images: Upload up to 4 reference images to guide the style and composition of generated SVGs
  • Control generation parameters: Adjust temperature, number of outputs, and token limits to fine-tune results
  • List available models: Query available QuiverAI models to discover supported operations and capabilities
  • Get clean SVG markup: Receive raw SVG content alongside downloadable files for easy embedding

In Sim, the Quiver integration enables your workflows to generate and vectorize graphics on demand. This is useful for creating dynamic illustrations, converting raster assets to scalable vectors, generating icons for applications, producing visual assets for content pipelines, or building design automation workflows. The generated SVGs are returned as files that can be passed to downstream blocks for further processing, storage, or delivery.

Usage Instructions

Generate SVG images from text prompts or vectorize raster images into SVGs using QuiverAI. Supports reference images, style instructions, and multiple output generation.

Tools

quiver_text_to_svg

Generate SVG images from text prompts using QuiverAI

Input

ParameterTypeRequiredDescription
apiKeystringYesQuiverAI API key
promptstringYesA text description of the desired SVG
modelstringYesThe model to use for SVG generation (e.g., "arrow-preview")
instructionsstringNoStyle or formatting guidance for the SVG output
referencesfileNoReference images to guide SVG generation (up to 4)
nnumberNoNumber of SVGs to generate (1-16, default 1)
temperaturenumberNoSampling temperature (0-2, default 1)
top_pnumberNoNucleus sampling probability (0-1, default 1)
max_output_tokensnumberNoMaximum output tokens (1-131072)
presence_penaltynumberNoToken penalty for prior output (-2 to 2, default 0)

Output

ParameterTypeDescription
successbooleanWhether the SVG generation succeeded
outputobjectGenerated SVG output
filefileGenerated SVG file
svgContentstringRaw SVG markup content
idstringGeneration request ID
usagejsonToken usage statistics
totalTokensnumberTotal tokens used
inputTokensnumberInput tokens used
outputTokensnumberOutput tokens used

quiver_image_to_svg

Convert raster images into vector SVG format using QuiverAI

Input

ParameterTypeRequiredDescription
apiKeystringYesQuiverAI API key
modelstringYesThe model to use for vectorization (e.g., "arrow-preview")
imagefileYesThe raster image to vectorize into SVG
temperaturenumberNoSampling temperature (0-2, default 1)
top_pnumberNoNucleus sampling probability (0-1, default 1)
max_output_tokensnumberNoMaximum output tokens (1-131072)
presence_penaltynumberNoToken penalty for prior output (-2 to 2, default 0)
auto_cropbooleanNoAutomatically crop the image before vectorizing
target_sizenumberNoSquare resize target in pixels (128-4096)

Output

ParameterTypeDescription
successbooleanWhether the vectorization succeeded
outputobjectVectorized SVG output
filefileGenerated SVG file
svgContentstringRaw SVG markup content
idstringVectorization request ID
usagejsonToken usage statistics
totalTokensnumberTotal tokens used
inputTokensnumberInput tokens used
outputTokensnumberOutput tokens used

quiver_list_models

List all available QuiverAI models

Input

ParameterTypeRequiredDescription
apiKeystringYesQuiverAI API key

Output

ParameterTypeDescription
successbooleanWhether the request succeeded
outputobjectAvailable models
modelsjsonList of available QuiverAI models
idstringModel identifier
namestringHuman-readable model name
descriptionstringModel capabilities summary
creatednumberUnix timestamp of creation
ownedBystringOrganization that owns the model
inputModalitiesjsonSupported input types (text, image, svg)
outputModalitiesjsonSupported output types (text, image, svg)
contextLengthnumberMaximum context window
maxOutputLengthnumberMaximum generation length
supportedOperationsjsonAvailable operations (svg_generate, svg_edit, svg_animate, svg_vectorize, chat_completions)
supportedSamplingParametersjsonSupported sampling parameters (temperature, top_p, top_k, repetition_penalty, presence_penalty, stop)

On this page

Start building today
Trusted by over 100,000 builders.
The open-source platform to build AI agents and run your agentic workforce.
Get started