MODELS.md

Higgsfield CLI Models

Generated from higgsfield model list and higgsfield model get <job_set_type>. Run those commands for the live schema.

Required flags are listed per model below. Media inputs (--image, --image-references, --start-image, --end-image, --video, --video-references, --audio, --audio-references) accept either a UUID (upload id or previous job id) or a local file path; paths are auto-uploaded.

Every flag accepts both spellings: --aspect_ratio and --aspect-ratio are equivalent.

Image (23)

cinematic_studio_2_5 — Cinematic Studio 2.5

flag	required	default	values
--aspect_ratio	false	1:1	1:1, 3:2, 2:3, 4:3, 3:4, 4:5, 5:4, 16:9, 9:16, 21:9
--batch_size	false	1	integer
--folder_id	false	—	string
--image-references (or --image) (0..14)	false	—	UUID or path
--mode	false	auto	string
--prompt	true	—	string
--resolution	false	1k	1k, 2k, 4k

Constraints:

• At most 14 image references are allowed.

flux_2 — FLUX.2

flag	required	default	values
--aspect_ratio	false	1:1	1:1, 4:3, 3:4, 16:9, 9:16
--image-references (or --image) (repeated)	false	—	UUID or path
--prompt	true	—	string
--resolution	false	1k	1k, 2k
--variant	false	pro	pro, flex, max

flux_kontext — Flux Kontext

flag	required	default	values
--aspect_ratio	false	1:1	1:1, 4:3, 3:4, 16:9, 9:16
--image-references (or --image) (0..4)	false	—	UUID or path
--prompt	true	—	string

Constraints:

• At most 4 image_references are allowed.

gpt_image_2 — GPT Image 2

High-fidelity image generation and editing for product photography, graphic design, and on-image text.

flag	required	default	values
--aspect_ratio	false	1:1	auto, 1:1, 4:3, 3:4, 16:9, 21:9, 9:16, 3:2, 2:3, 4:5, 5:4
--background	false	—	auto, opaque, transparent
--image-references (or --image) (repeated)	false	—	UUID or path
--prompt	true	—	string
--quality	false	high	low, medium, high
--resolution	false	2k	1k, 2k, 4k
--is_inpaint	false	false	boolean; requires a mask and at least one reference image
--mask	false	—	media reference object; use with --is_inpaint true

higgsfield generate create gpt_image_2 --prompt "A studio product photo" --wait

gpt_image_2_5 — GPT Image 2.5

Recommended default for high-fidelity image generation, design, and on-image text. Supports reference-guided editing with Flare and Sunburst variants and higher quality tiers for final assets.

flag	required	default	values
--aspect_ratio	false	1:1	auto, 1:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16, 21:9, 27:16, 16:27, 9:8, 8:9, 4:5, 5:4
--background	false	—	auto, opaque, transparent
--image-references (or --image) (repeated)	false	—	UUID or path; at most 16
--prompt	true	—	string
--quality	false	low	low, medium, high, xhigh, max
--resolution	false	1k	1k, 2k, 4k
--variant	false	flare	flare, sunburst

higgsfield generate create gpt_image_2_5 --prompt "A landscape product photo" --resolution 2k --wait

grok_image — Grok Image

flag	required	default	values
--aspect_ratio	false	1:1	1:1, auto, 1:2, 2:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16
--image-references (or --image) (repeated)	false	—	UUID or path
--mode	false	std	std, quality
--prompt	true	—	string
--resolution	false	1k	1k, 2k

image_auto — Image Auto

flag	required	default	values
--aspect_ratio	false	1:1	1:1, 4:3, 3:4, 16:9, 9:16
--image-references (or --image) (0..14)	false	—	UUID or path
--prompt	true	—	string

Constraints:

• At most 14 image references are allowed.

image_background_remover — Image Background Remover

flag	required	default	values
--image-references (or --image) (repeated)	false	—	UUID or path

Constraints:

• Exactly one image_references entry is required.

Example:

higgsfield generate create image_background_remover --image ./image.png --wait

kling_omni_image — Kling O1 Image

flag	required	default	values
--aspect_ratio	false	1:1	1:1, auto, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 21:9
--image-references (or --image) (0..10)	false	—	UUID or path
--prompt	true	—	string
--resolution	false	1k	1k, 2k

Constraints:

• At most 10 image_references are allowed.
• Aspect_ratio 'auto' requires at least one image_reference.

marketing_studio_image — Marketing Studio Image

flag	required	default	values
--aspect_ratio	false	1:1	auto, 1:1, 3:2, 2:3, 4:3, 3:4, 4:5, 5:4, 9:16, 16:9, 21:9
--image-references (or --image) (0..14)	false	—	UUID or path
--prompt	true	—	string
--resolution	false	1k	1k, 2k, 4k

Constraints:

• At most 14 image_references are allowed.
• Aspect_ratio 'auto' requires at least one image_reference.
• 'prompt' or at least one image_reference is required.

nano_banana — Nano Banana

flag	required	default	values
--aspect_ratio	false	1:1	1:1, 3:2, 2:3, 4:3, 3:4, 4:5, 5:4, 9:16, 16:9, 21:9
--image-references (or --image) (0..8)	false	—	UUID or path
--prompt	true	—	string

Constraints:

• At most 8 image_references are allowed.

nano_banana_2 — Nano Banana Pro

flag	required	default	values
--aspect_ratio	false	1:1	1:1, 3:2, 2:3, 4:3, 3:4, 4:5, 5:4, 9:16, 16:9, 21:9
--image-references (or --image) (0..14)	false	—	UUID or path
--prompt	true	—	string
--resolution	false	2k	1k, 2k, 4k

Constraints:

• A non-empty prompt or at least one image reference is required.
• At most 14 image references are allowed.

nano_banana_2_lite — Nano Banana 2 Lite

flag	required	default	values
--aspect_ratio	false	1:1	auto, 1:1, 3:2, 2:3, 4:3, 3:4, 4:5, 5:4, 9:16, 16:9, 21:9
--image-references (or --image) (0..14)	false	—	UUID or path
--prompt	true	—	string
--resolution	false	1k	1k
--thinking	false	HIGH	MINIMAL, HIGH

Constraints:

• A non-empty prompt or at least one image reference is required.
• At most 14 image references are allowed.
• Aspect_ratio 'auto' requires at least one image reference.

Example:

higgsfield generate create nano_banana_2_lite --prompt "clean product sketch, white background" --image ./reference.png --resolution 1k --wait

nano_banana_flash — Nano Banana 2

flag	required	default	values
--aspect_ratio	false	1:1	1:1, 3:2, 2:3, 4:3, 3:4, 4:5, 5:4, 9:16, 16:9, 21:9
--image-references (or --image) (repeated)	false	—	UUID or path
--prompt	true	—	string
--resolution	false	1k	1k, 2k, 4k

openai_hazel — OpenAI Hazel

flag	required	default	values
--aspect_ratio	false	1:1	1:1, 3:2, 2:3, auto
--image-references (or --image) (0..16)	false	—	UUID or path
--prompt	true	—	string
--quality	false	medium	low, medium, high

Constraints:

• Aspect_ratio 'auto' requires at least one image reference.
• At most 16 image references are allowed.

outpaint — Outpaint

flag	required	default	values
--aspect_ratio	false	21:9	auto, 1:1, 3:2, 2:3, 4:3, 3:4, 4:5, 5:4, 9:16, 16:9, 21:9
--folder_id	false	—	string
--image-references (or --image) (repeated)	false	—	UUID or path

Constraints:

• Exactly one image_references entry is required.

Example:

higgsfield generate create outpaint --image ./image.png --aspect_ratio 16:9 --wait

recraft_v4_1 — Recraft V4.1

flag	required	default	values
--aspect_ratio	false	1:1	1:1, 3:4, 4:3, 4:5, 5:4, 3:2, 2:3, 16:9, 9:16
--background_color	false	—	string
--colors	false	—	array
--model_type	false	standard	standard, vector, utility, utility_vector
--prompt	true	—	string
--resolution	false	1k	1k, 2k

Constraints:

• Each color in colors must be a #RRGGBB hex string (six hex digits, no alpha).
• Background_color must be a #RRGGBB hex string (six hex digits, no alpha) or null.

Example:

higgsfield generate create recraft_v4_1 --prompt "minimal vector logo mark for a coffee brand" --model_type vector --resolution 2k --wait

seedream_v4_5 — Seedream 4.5

flag	required	default	values
--aspect_ratio	false	1:1	1:1, 4:3, 16:9, 3:2, 21:9, 3:4, 9:16, 2:3
--image-references (or --image) (0..14)	false	—	UUID or path
--prompt	true	—	string
--quality	false	basic	basic, high

Constraints:

• At most 14 image_references are allowed.

seedream_v5_lite — Seedream V5 Lite

flag	required	default	values
--aspect_ratio	false	1:1	1:1, 4:3, 3:4, 16:9, 9:16
--image-references (or --image) (repeated)	false	—	UUID or path
--prompt	true	—	string
--quality	false	basic	basic, high

soul_cast — Soul Cast

flag	required	default	values
--aspect_ratio	true	—	16:9
--budget	false	50	integer
--prompt	false	—	string

soul_cinematic — Soul Cinematic

flag	required	default	values
--aspect_ratio	false	1:1	1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 21:9
--image-references (or --image) (single)	false	—	UUID or path
--prompt	true	—	string
--quality	false	2k	1.5k, 2k
--soul-id (or --custom-reference-id)	false	—	Soul UUID

Constraints:

• At most one image reference is allowed.

soul_location — Soul Location

flag	required	default	values
--aspect_ratio	false	1:1	1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 21:9, 9:21
--prompt	true	—	string

text2image_soul_v2 — Higgsfield Soul V2

flag	required	default	values
--aspect_ratio	false	1:1	1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3
--image-references (or --image) (single)	false	—	UUID or path
--prompt	true	—	string
--quality	false	2k	1.5k, 2k
--soul-id (or --custom-reference-id)	false	—	Soul UUID

Constraints:

• At most one image reference is allowed.

z_image — Z Image

flag	required	default	values
--aspect_ratio	false	1:1	1:1, 4:3, 3:4, 16:9, 9:16
--prompt	true	—	string

Video (22)

brain_activity — Virality Predictor

flag	required	default	values
--folder_id	false	—	string
--video (single)	true	—	UUID or path

Analyzes a video and predicts audience engagement. Pass the video with --video.

cinematic_studio_3_0 — Cinematic Studio 3.0

flag	required	default	values
--aspect_ratio	false	16:9	auto, 21:9, 16:9, 4:3, 1:1, 3:4, 9:16
--audio-references (or --audio) (repeated)	false	—	UUID or path
--batch_size	false	1	integer
--duration	false	5	integer
--end-image (single)	false	—	UUID or path
--enhance_prompt	false	false	boolean
--folder_id	false	—	string
--generate_audio	false	false	boolean
--genre	false	auto	auto, action, horror, comedy, noir, drama, epic
--image-references (or --image) (repeated)	false	—	UUID or path
--multi_prompt	false	—	array
--multi_shot_mode	false	custom	auto, custom
--multi_shots	false	false	boolean
--preset_id	false	—	string
--prompt	true	—	string
--prompt_language	false	zh	en, zh
--resolution	false	720p	480p, 720p, 1080p, 4k
--speedramp	false	auto	auto, linear, slowmo, speedup, fast_to_slowmo, slowmo_to_fast, super_slowmo, impact
--start-image (single)	false	—	UUID or path
--video-references (or --video) (repeated)	false	—	UUID or path

Constraints:

• At most 15 media references are allowed in total.

cinematic_studio_video — Cinematic Studio Video

flag	required	default	values
--aspect_ratio	false	16:9	1:1, 4:3, 3:4, 16:9, 9:16
--duration	false	5	5, 10
--end-image (single)	false	—	UUID or path
--prompt	true	—	string
--slow_motion	false	false	boolean
--sound	false	true	boolean
--start-image (single)	false	—	UUID or path

Constraints:

• End_image requires start_image to also be provided.

cinematic_studio_video_3_5 — Cinematic Studio Video 3.5

flag	required	default	values
--aspect_ratio	false	auto	auto, 21:9, 16:9, 4:3, 1:1, 3:4, 9:16
--audio-references (or --audio) (0..15)	false	—	UUID or path
--camera_style	false	—	classic_static, silent_machine, one_take, epic_scale, intimate_observer, impossible_camera, documentary_snap, raw_chaos, dreamy_flow
--color_grading	false	—	naturalistic_clean, bleached_warm, hyper_neon, teal_orange_epic, sodium_decay, cold_steel, bleach_bypass, classic_bw
--duration	false	15	integer
--end-image (single)	false	—	UUID or path
--enhance_prompt	false	false	boolean
--generate_audio	false	false	boolean
--genre	false	auto	auto, action, horror, comedy, noir, drama, epic
--image-references (or --image) (repeated)	false	—	UUID or path
--light_scheme	false	—	soft_cross, contre_jour, overhead_fall, window, practicals, silhouette
--multi_prompt	false	—	array
--multi_shot_mode	false	custom	auto, custom
--multi_shots	false	false	boolean
--prompt	false	—	string
--prompt_language	false	zh	en, zh
--resolution	false	720p	480p, 720p, 1080p
--start-image (single)	false	—	UUID or path
--style_prompt	false	—	string
--video-references (or --video) (repeated)	false	—	UUID or path

Constraints:

• At most 15 media references are allowed in total (image_references + start_image + end_image + video_references + audio_references).
• Inline style axes (camera_style/light_scheme/color_grading) and style_prompt are mutually exclusive.

cinematic_studio_video_v2 — Cinematic Studio Video V2

flag	required	default	values
--aspect_ratio	false	16:9	1:1, 4:3, 3:4, 16:9, 9:16
--batch_size	false	1	integer
--cfg_scale	false	0.5	number
--duration	false	5	integer
--end-image (single)	false	—	UUID or path
--folder_id	false	—	string
--genre	false	auto	auto, action, horror, comedy, western, suspense, intimate, spectacle
--image-references (or --image) (repeated)	false	—	UUID or path
--kling_element_ids	false	—	array
--mode	false	std	pro, std
--multi_prompt	false	—	array
--multi_shot_mode	false	custom	auto, custom
--multi_shots	false	false	boolean
--preset_id	false	—	string
--prompt	true	—	string
--sound	false	on	on, off
--speedramp	false	auto	auto, custom, linear, slowmo, speedup, impact
--start-image (single)	false	—	UUID or path

gemini_omni — Gemini Omni Flash

flag	required	default	values
--aspect_ratio	false	16:9	16:9, 9:16
--duration	false	8	4, 6, 8, 10
--image-references (or --image) (0..7)	false	—	UUID or path
--prompt	true	—	string
--resolution	false	720p	720p
--video-references (or --video) (single)	false	—	UUID or path

Constraints:

• At most 1 video_references entry is allowed.
• When a video reference is provided, at most 5 image_references are allowed.
• At most 7 image_references are allowed.

Example:

higgsfield generate create gemini_omni --prompt "cinematic product reveal with smooth camera motion" --image ./product.png --duration 8 --resolution 720p --wait

grok_video — Grok Video

flag	required	default	values
--aspect_ratio	false	16:9	16:9, 9:16, 1:1
--duration	false	5	integer
--prompt	true	—	string
--start-image (single)	false	—	UUID or path

grok_video_v15 — Grok Video 1.5

flag	required	default	values
--duration	false	5	integer
--prompt	true	—	string
--resolution	false	720p	480p, 720p
--start-image (single)	true	—	UUID or path

Example:

higgsfield generate create grok_video_v15 --prompt "cinematic handheld shot, neon rainy street" --start-image ./image.png --duration 5 --resolution 720p --wait

kling2_6 — Kling 2.6 Video

flag	required	default	values
--aspect_ratio	false	16:9	16:9, 9:16, 1:1
--duration	false	5	5, 10
--prompt	true	—	string
--sound	false	true	boolean
--start-image (single)	false	—	UUID or path

kling3_0 — Kling v3.0

flag	required	default	values
--aspect_ratio	false	16:9	16:9, 9:16, 1:1
--duration	false	5	integer
--end-image (single)	false	—	UUID or path
--mode	false	std	std, pro, 4k
--prompt	true	—	string
--sound	false	on	on, off
--start-image (single)	false	—	UUID or path

kling3_0_turbo — Kling 3.0 Turbo

flag	required	default	values
--aspect_ratio	false	16:9	16:9, 9:16, 1:1
--duration	false	5	integer
--prompt	true	—	string
--resolution	false	720p	720p, 1080p
--start-image (single)	false	—	UUID or path

Example:

higgsfield generate create kling3_0_turbo --prompt "fast handheld product reveal on a clean studio table" --start-image ./first.png --duration 5 --resolution 720p --wait

marketing_studio_video — Marketing Studio Video

flag	required	default	values
--ad_reference_id	false	—	ad-reference UUID
--aspect_ratio	false	16:9	auto, 21:9, 16:9, 4:3, 1:1, 3:4, 9:16
--avatar_ids	false	—	array
--avatars	false	—	array
--duration	false	15	integer
--end-image (single)	false	—	UUID or path
--generate_audio	false	false	boolean
--hook_id	false	—	hook UUID
--image-references (or --image) (repeated)	false	—	UUID or path
--mode	false	ugc	string
--product_ids	false	—	array
--prompt	true	—	string
--resolution	false	720p	480p, 720p, 1080p
--setting_id	false	—	setting UUID
--specific_mode	false	default	default, web_product, from_storyboard
--start-image (single)	false	—	UUID or path
--storyboard_id	false	—	storyboard UUID
--web_product_ids	false	—	array
--web_product_type	false	—	desktop, mobile

Constraints:

• Ad_reference_id cannot be combined with hook_id or setting_id.
• Product_ids and web_product_ids cannot both be set.

minimax_hailuo — Minimax Hailuo

flag	required	default	values
--duration	false	6	6, 10
--end-image (single)	false	—	UUID or path
--prompt	true	—	string
--resolution	false	768	512, 768, 1080
--start-image (single)	false	—	UUID or path
--variant	false	minimax-2.3	minimax, minimax-fast, minimax-2.3, minimax-2.3-fast

Constraints:

• Resolution '512' is incompatible with end_image.
• Resolution '512' is not supported for 'minimax-2.3' or 'minimax-2.3-fast'.
• Resolution '1080' is not available for 10 second duration.
• Start_image or end_image is required unless variant is 'minimax-2.3'.
• End_image is not supported for 'minimax-2.3' or 'minimax-2.3-fast'.

seedance1_5 — Seedance 1.5 Pro

flag	required	default	values
--aspect_ratio	false	16:9	auto, 16:9, 9:16, 4:3, 3:4, 1:1, 21:9
--duration	false	4	4, 8, 12
--end-image (single)	false	—	UUID or path
--generate_audio	false	true	boolean
--prompt	true	—	string
--resolution	false	720p	480p, 720p, 1080p
--start-image (single)	false	—	UUID or path

Constraints:

• End_image requires start_image to be set.

seedance_2_0 — Seedance 2.0

flag	required	default	values
--aspect_ratio	false	16:9	auto, 16:9, 9:16, 4:3, 3:4, 1:1, 21:9
--audio-references (or --audio) (0..3)	false	—	UUID or path
--bitrate_mode	false	standard	standard, high
--duration	false	5	integer
--end-image (single)	false	—	UUID or path
--generate_audio	false	true	boolean
--genre	false	auto	auto, action, horror, comedy, noir, drama, epic
--image-references (or --image) (repeated)	false	—	UUID or path
--mode	false	std	std, fast
--prompt	true	—	string
--resolution	false	720p	480p, 720p, 1080p, 4k
--start-image (single)	false	—	UUID or path
--video-references (or --video) (0..3)	false	—	UUID or path

Constraints:

• At most 9 image references are allowed (counting start_image and end_image).
• At most 3 video_references are allowed.
• At most 3 audio_references are allowed.
• At most 12 reference files are allowed in total across images, videos, and audios (including start_image and end_image).
• Audio_references require at least one image, video, start_image, or end_image.
• Mode 'fast' supports only 480p/720p; use mode 'std' for 1080p/4k.

seedance_2_0_mini — Seedance 2.0 Mini

flag	required	default	values
--aspect_ratio	false	16:9	auto, 16:9, 9:16, 4:3, 3:4, 1:1, 21:9
--audio-references (or --audio) (0..3)	false	—	UUID or path
--bitrate_mode	false	standard	standard, high
--duration	false	5	integer
--end-image (single)	false	—	UUID or path
--generate_audio	false	true	boolean
--genre	false	auto	auto, action, horror, comedy, noir, drama, epic
--image-references (or --image) (repeated)	false	—	UUID or path
--prompt	true	—	string
--resolution	false	720p	480p, 720p
--start-image (single)	false	—	UUID or path
--video-references (or --video) (0..3)	false	—	UUID or path

Constraints:

• At most 9 image references are allowed (counting start_image and end_image).
• At most 3 video_references are allowed.
• At most 3 audio_references are allowed.
• At most 12 reference files are allowed in total across images, videos, and audios (including start_image and end_image).
• Audio_references require at least one image, video, start_image, or end_image.

veo3 — Google Veo 3

flag	required	default	values
--aspect_ratio	false	16:9	16:9, 9:16
--prompt	true	—	string
--start-image (single)	true	—	UUID or path
--variant	false	veo-3-fast	veo-3-preview, veo-3-fast

veo3_1 — Google Veo 3.1

flag	required	default	values
--aspect_ratio	false	16:9	16:9, 9:16
--duration	false	8	4, 6, 8
--prompt	true	—	string
--quality	false	basic	basic, high, ultra
--start-image (single)	false	—	UUID or path
--variant	false	veo-3-1-fast	veo-3-1-preview, veo-3-1-fast

veo3_1_lite — Google Veo 3.1 Lite

flag	required	default	values
--aspect_ratio	false	16:9	16:9, 9:16, auto
--duration	false	8	4, 6, 8
--end-image (single)	false	—	UUID or path
--generate_audio	false	false	boolean
--prompt	true	—	string
--start-image (single)	false	—	UUID or path

Constraints:

• Duration must be 8 when both start_image and end_image are set.

video_background_remover — Video Background Remover

flag	required	default	values
--video-references (or --video) (repeated)	false	—	UUID or path

Constraints:

• Exactly one video is required in video_references.

Example:

higgsfield generate create video_background_remover --video ./video.mp4 --wait

wan2_6 — Wan 2.6 Video

flag	required	default	values
--aspect_ratio	false	16:9	16:9, 9:16, 1:1
--audio-references (or --audio) (repeated)	false	—	UUID or path
--duration	false	5	5, 10, 15
--image-references (or --image) (repeated)	false	—	UUID or path
--prompt	true	—	string
--quality	false	720p	720p, 1080p
--video-references (or --video) (repeated)	false	—	UUID or path

Constraints:

• Reference-to-video (with video_references) supports only 5 or 10 second durations.

wan2_7 — Wan 2.7

flag	required	default	values
--aspect_ratio	false	16:9	16:9, 9:16, 1:1, 4:3, 3:4
--audio-references (or --audio) (single)	false	—	UUID or path
--duration	false	5	integer
--end-image (single)	false	—	UUID or path
--prompt	true	—	string
--resolution	false	720p	720p, 1080p
--start-image (single)	false	—	UUID or path

Constraints:

• End_image requires start_image.
• At most 1 audio reference is allowed.

Video explainer jobs

These job types have similar names but different responsibilities:

• explainer_video assembles ordered video/audio block pairs and is the final step used by the public explainer skill.
• video_explainer is an alternate monolithic server-run workflow. The public skill does not use it.

video_explainer — Explainer Video

flag	required	default	values
--prompt	conditional	—	string, max 5000 characters
--medias	false	[]	JSON array via @file; up to 14 images/files
--aspect_ratio	false	16:9	16:9, 9:16
--duration	true	—	integer, 20–600, multiple of 10
--preset_id	false	—	UUID from higgsfield preset list video-explainer
--voice_type	false	—	preset, element
--voice_id	false	—	voice ID from higgsfield voices list
--folder_id	false	—	UUID

Constraints:

• prompt is required unless at least one image/file is attached.
• Video and audio attachments are rejected.
• voice_type and voice_id must be provided together.
• preset_id controls style only; it does not replace the topic.
• Job creation resolves the preset and imports its hidden style image automatically.

Examples:

higgsfield preset list video-explainer --json

higgsfield generate create video_explainer \
  --prompt "Explain compound interest to teenagers. Narration language: English." \
  --duration 60 \
  --aspect_ratio 16:9 \
  --preset_id <preset_id> \
  --wait

higgsfield generate create video_explainer \
  --prompt "Explain the supplied diagrams. Narration language: Spanish." \
  --image ./diagram-1.png \
  --image ./diagram-2.png \
  --duration 90 \
  --aspect_ratio 9:16 \
  --voice_type preset \
  --voice_id <voice_id> \
  --wait

This job remains available as a direct API surface, but higgsfield-video-explainer uses separately generated Seed Audio and Gemini Omni blocks followed by explainer_video so every narration take maps explicitly to one clip.

explainer_video — Explainer Video Assembler

Assembler for callers that already generated ordered clip/audio pairs. It does not plan a narrative, choose a style, generate clips, or create narration.

flag	required	default	values
--width	true	—	even integer greater than 1
--height	true	—	even integer greater than 1
--items	true	—	JSON array via @file; at least 2 items
--subtitles	false	—	JSON object; font: patrick, caveat, marker, anton

Each item contains video: {id,type} and optional audio: {id,type}. Completed generation jobs may use the generic types video_job and audio_job:

[
  {
    "video": {"id": "<clip-1-job-id>", "type": "video_job"},
    "audio": {"id": "<voice-1-job-id>", "type": "audio_job"}
  },
  {
    "video": {"id": "<clip-2-job-id>", "type": "video_job"},
    "audio": {"id": "<voice-2-job-id>", "type": "audio_job"}
  }
]

higgsfield generate create explainer_video \
  --items @blocks.json \
  --width 1280 \
  --height 720 \
  --subtitles '{"font":"patrick"}' \
  --wait

3D (5)

3d_rigging — 3D Rigging

flag	required	default	values
--animation_action_id	false	—	integer
--enable_animation	false	false	boolean
--enable_safety_checker	false	—	boolean
--folder_id	false	—	string
--height_meters	false	—	number
--model_url	true	—	string

image_to_3d — Image to 3D

flag	required	default	values
--animation_action_id	false	—	integer
--enable_animation	false	false	boolean
--enable_pbr	false	—	boolean
--enable_rigging	false	false	boolean
--enable_safety_checker	false	—	boolean
--folder_id	false	—	string
--image-references (or --image) (repeated)	true	—	UUID or path
--pose_mode	false	—	a-pose, t-pose
--rigging_height_meters	false	—	number
--seed	false	—	integer
--should_remesh	false	—	boolean
--should_texture	false	false	boolean
--symmetry_mode	false	—	off, auto, on
--target_polycount	false	—	integer
--texture_image_url	false	—	string
--texture_prompt	false	—	string
--topology	false	—	quad, triangle

Constraints:

• Enable_animation requires enable_rigging=true.
• Animation_action_id is required when enable_animation=true.
• Texture_prompt requires should_texture=true.
• Texture_image_url requires should_texture=true.
• Enable_pbr requires should_texture=true.

multi_image_to_3d — Multi-Image to 3D

flag	required	default	values
--animation_action_id	false	—	integer
--enable_animation	false	false	boolean
--enable_pbr	false	—	boolean
--enable_rigging	false	false	boolean
--enable_safety_checker	false	—	boolean
--folder_id	false	—	string
--image-references (or --image) (repeated)	true	—	UUID or path
--pose_mode	false	—	a-pose, t-pose
--rigging_height_meters	false	—	number
--seed	false	—	integer
--should_remesh	false	—	boolean
--should_texture	false	false	boolean
--symmetry_mode	false	—	off, auto, on
--target_polycount	false	—	integer
--texture_image_url	false	—	string
--texture_prompt	false	—	string
--topology	false	—	quad, triangle

Constraints:

• Enable_animation requires enable_rigging=true.
• Animation_action_id is required when enable_animation=true.
• Texture_prompt requires should_texture=true.
• Texture_image_url requires should_texture=true.
• Enable_pbr requires should_texture=true.

Example:

higgsfield generate create multi_image_to_3d --image ./front.png --image ./side.png --should_texture true --wait

sam_3_3d — 3D Objects

flag	required	default	values
--detection_threshold	false	—	number
--export_textured_glb	false	true	boolean
--folder_id	false	—	string
--image-references (or --image) (repeated)	true	—	UUID or path
--prompt	false	—	string
--seed	false	—	integer

tripo_3d — Text to 3D

flag	required	default	values
--auto_size	false	false	boolean
--face_limit	false	—	integer
--geometry_quality	false	standard	standard, detailed
--negative_prompt	false	—	string
--pbr	false	true	boolean
--prompt	true	—	string
--texture	false	true	boolean
--texture_quality	false	standard	standard, detailed

Audio (5)

inworld_text_to_speech — Inworld Text to Speech

flag	required	default	values
--prompt	true	—	string
--voice	true	—	113 built-in voices — run higgsfield model get inworld_text_to_speech for the list

mirelo_text_to_audio — Mirelo Text to Audio

flag	required	default	values
--duration	true	—	number
--prompt	true	—	string

Example:

higgsfield generate create mirelo_text_to_audio --prompt "glass breaking in a large hall" --duration 4 --wait

seed_audio — Seed Audio 1.0

flag	required	default	values
--audio-references (or --audio) (0..2)	false	—	UUID or path
--format	false	wav	wav, mp3, pcm, ogg_opus
--image-references (or --image) (repeated)	false	—	UUID or path
--loudness_rate	false	0	integer
--pitch_rate	false	0	integer
--prompt	true	—	string
--sample_rate	false	24000	8000, 16000, 24000, 32000, 44100, 48000
--speech_rate	false	0	integer
--voice_id	false	—	string
--voice_type	false	—	preset, element

For a cloned or preset voice, take --voice_id and --voice_type from higgsfield voices list.

Constraints:

• Image_references and audio_references are mutually exclusive.
• Voice_type and voice_id must be provided together.
• A voice cannot be combined with an image reference.
• A voice allows at most 2 additional audio references.

Example:

higgsfield generate create seed_audio --prompt "Welcome to the show!" --voice_type preset --voice_id <voice_id> --wait

sonilo_music — Sonilo Music

flag	required	default	values
--duration	true	—	number
--prompt	true	—	string

Example:

higgsfield generate create sonilo_music --prompt "cinematic synthwave track" --duration 12 --wait

text2speech_v2 — Text to Speech

flag	required	default	values
--prompt	true	—	string
--variant	true	—	elevenlabs, minimax, seed_speech, vibe_voice, cozy_voice
--voice_id	true	—	string
--voice_type	true	—	preset, element

Discover voices (--voice_id / --voice_type) with higgsfield voices list.

Constraints:

• Prompt character limit depends on the engine: elevenlabs/vibe_voice/cozy_voice 5000, minimax 10000, seed_speech 15000.

Example:

higgsfield generate create text2speech_v2 --prompt "Hello from Higgsfield" --variant elevenlabs --voice_type preset --voice_id <voice_id> --wait

Additional utility job types (upscale, background removal, transcription, etc.) are available; run higgsfield model get <job_type> for their schemas: autosprite, bytedance_image_upscale, bytedance_video_upscale, clipify, color_grading_lut, llm_text, sam_3_video, speech2text, topaz_image, topaz_video, video_deflicker, video_upscale.