Documentation
Models, capabilities and parameters
Read DesignCat’s public catalog and use authenticated discovery to obtain the exact schema for your task.
A model can have several operations
Text-to-image, image-to-image, image-to-video and video editing are different contracts. A capability determines the required inputs, supported references, resolution, duration and output. Parameter ranges can also have cross-field restrictions that the server checks when you submit.
Public catalog
The public JSON snapshot includes model identity, capabilities, parameter summaries, surface mappings and pricing-source metadata. It contains no account resources or credentials. Use the linked model pages for readable tables.
https://designcat.io/data/catalog.json
Authenticated discovery
Agents should use list_agent_models, get_agent_model_schema and
get_agent_catalog on their authenticated connection. The HTTP counterparts
are under /api/v1/agent/. These live contracts are authoritative for a
submission; a published snapshot describes a particular release.
Language-model requests use the Responses API separately from media generation. Their context windows and input modalities do not establish availability in every product surface.
Credits and evidence
Inspect the current credit estimate before generation. Editorial quality scores and hypothetical cost normalizations are different from recorded study results.