List Available AI Models
Retrieve the list of AI models available through the Hedra API, including image, video, and audio generation models.
Authorizations
Query Parameters
Generation type enum
NOTE: this enum is used to determine the type of generation and is used to determine the type of asset that will be generated.
image, video, text_to_speech, speech_to_speech, voice_clone, audio_isolation, video_stitching, video_upscale, video_to_video, image_upscale, agent_response, audio_from_video, text_to_sound, assets_to_image_text_prompt, assets_to_audio_text_prompt Response
Successful Response
Stable cross-environment identifier for the model, e.g. google/nano-banana. Unique and identical across local/staging/production. Prefer slug over id when referencing models — id is environment-specific and will be removed in a later migration (model-registry plan step 7).
Name of the model
Description of the model.
Type of generation the model applies to.
Pricing details of the model.
Environment-specific UUID of the model, or null for a code-backed model with no ai_models row (identified by slug). Being removed in favor of slug (model-registry plan step 7); prefer slug.
Aspect ratios the model supports.
If set to (min, max), the model accepts any keyframe whose width/height ratio is in [min, max], not just the values in aspect_ratios. The model snaps the output to its internal grid. aspect_ratios then serves as UI/preset labels rather than as an enum gate.
Resolutions the model supports.
Backend-declared default output resolution for this model.
Durations the model supports.
Whether the model is conditioned by a start frame.
Whether the model is conditioned by an end frame.
Whether the model is conditioned by audio input.
Whether the model requires video input (video-to-video).
Whether the model requires character orientation (motion control).
Average generation duration in milliseconds.
Tags for model categorization.
Maximum output duration in milliseconds for video/audio models.
Whether the model supports custom resolution.
Extensible pricing information with dimension modifiers for resolution, audio, etc.
Width and height for each aspect_ratio and resolution tuple.
Minimum character count for text prompts. Null means no minimum.
Maximum character count for text prompts. Null means no maximum.
List of input modes the model supports. Each mode groups mutually exclusive input slots. The frontend picks one mode. text_to_video (no inputs) is always implicitly available for VIDEO type models. Null means the model has no declarative input specifications (use requires_* booleans).
URL of the model's logo in SVG format.
Whether this is a premium model.
Display order for UI sorting. Lower values appear first.