> ## Documentation Index
> Fetch the complete documentation index at: https://composio-preview.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Higgsfield

> Higgsfield AI provides asynchronous image and video generation APIs powered by multiple creative AI models.

![Higgsfield logo](https://mintcdn.com/composio-preview/o1IUBrrsmtgkhmyf/images/toolkits/higgsfield.svg?fit=max\&auto=format\&n=o1IUBrrsmtgkhmyf\&q=85\&s=a282c74984fca24ac023b7c7ee19f2a8)

* **Category:** media generation
* **Auth:** API\_KEY
* **Composio-managed OAuth available?** N/A
* **Tools:** 14
* **Triggers:** 0
* **Slug:** `HIGGSFIELD`
* **Version:** 20261101\_00
* **Instant:** Supported on the latest version. Support varies by tool.

## Tools

> [accordion] Cancel Request
>
> **Slug:** `HIGGSFIELD_CANCEL_REQUEST`
>
> Cancel a Higgsfield generation request while it is still queued; requests already processing or finished cannot be canceled. Managed generation requests cannot be canceled.
>
> #### Input Parameters
>
> **request\_id** `string` *required*
>
> Queued generation request ID returned by a generation action.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Create Character Reference
>
> **Slug:** `HIGGSFIELD_CREATE_CHARACTER_REFERENCE`
>
> Start training a reusable Higgsfield Soul ID character reference from one or more public images using your connected account. The returned ID can be used to check training status.
>
> #### Input Parameters
>
> **name** `string` *required*
>
> Name of the character reference (maximum 100 characters).
>
> **input\_images** `array` *required*
>
> One to 100 public images used to train the character reference.
>
> **model\_version** `string ("v1" | "v2" | "cinema")`
>
> SOUL model family that will use the character reference.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Estimate Soul V2 Cost
>
> **Slug:** `HIGGSFIELD_ESTIMATE_SOUL_V2_COST`
>
> Estimate your connected account's credit and USD cost for a Higgsfield Soul V2 Standard image request without starting generation.
>
> #### Input Parameters
>
> **seed** `integer`
>
> Optional deterministic generation seed from 1 to 1000000. Omit it to let Higgsfield choose a random seed.
>
> **prompt** `string` *required*
>
> Text instructions for the image generation request to estimate.
>
> **style\_id** `string`
>
> Identifier of a style supported by Soul V2.
>
> **batch\_size** `integer ("1" | "4")`
>
> Number of output images. Soul V2 supports only 1 or 4.
>
> **resolution** `string ("720p" | "1080p")`
>
> Output resolution tier.
>
> **aspect\_ratio** `string ("9:16" | "16:9" | "4:3" | "3:4" | "1:1" | "2:3" | "3:2")`
>
> Output width-to-height ratio.
>
> **enhance\_prompt** `boolean`
>
> Whether Higgsfield should enhance the prompt.
>
> **style\_strength** `number`
>
> Style strength from 0 to 1. Higgsfield currently accepts this field without applying it to generation.
>
> **custom\_reference\_id** `string`
>
> Optional identifier of a completed character reference owned by the connected account.
>
> **custom\_reference\_strength** `number`
>
> Character-reference strength from 0 to 1. When custom\_reference\_id is set, this must be greater than 0.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Generate Content (Deprecated)
>
> **Slug:** `HIGGSFIELD_GENERATE_CONTENT_BY_MODEL_ID`
>
> Submit a generation request using an allowed Higgsfield model route and your connected account, returning a request ID for status polling. This fallback is deprecated.
>
> #### Input Parameters
>
> **prompt** `string` *required*
>
> The text prompt describing what to generate
>
> **duration** `integer`
>
> Duration in seconds for video generation models
>
> **model\_id** `string ("reve/text-to-image")` *required*
>
> Deprecated fallback model route not covered by a typed action. Existing legacy model IDs remain accepted for compatibility but are not advertised.
>
> **image\_url** `string`
>
> Public input image URL for image-to-video or image-to-image models
>
> **resolution** `string`
>
> Provider-supported output resolution
>
> **webhook\_url** `string`
>
> Public HTTPS URL for notification when generation finishes
>
> **aspect\_ratio** `string`
>
> Aspect ratio for the output, such as 16:9, 1:1, 9:16, or 4:3
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Generate Image With References
>
> **Slug:** `HIGGSFIELD_GENERATE_IMAGE_WITH_REFERENCES`
>
> Submit a Higgsfield Popcorn request that uses one or more reference images and return the request ID used to track the asynchronous result. Managed estimates above \$2 are rejected before generation submission. No polling is performed; use the returned request ID with GET\_REQUEST\_STATUS. Managed generation charges the upfront estimate when submission is accepted, including if generation later fails. Managed requests cannot be canceled.
>
> **Instant:** Supported on the latest version.
>
> **Instant pricing:** Popcorn Auto: varies by references, count and resolution; \$2 max per request.
>
> #### Input Parameters
>
> **seed** `integer`
>
> Optional deterministic generation seed (1-1000000).
>
> **prompt** `string` *required*
>
> Text description of the image to generate using the references.
>
> **image\_urls** `array` *required*
>
> Public URLs of 1-8 reference images. Each image must already be publicly accessible over HTTP or HTTPS.
>
> **num\_images** `integer`
>
> Number of images to generate (1-8).
>
> **resolution** `string ("720p" | "1600p")`
>
> Output quality.
>
> **webhook\_url** `string`
>
> Optional public HTTPS URL that receives the terminal completed, failed, or nsfw payload.
>
> **aspect\_ratio** `string ("1:1" | "4:3" | "3:4" | "3:2" | "2:3" | "16:9" | "9:16")`
>
> Aspect ratio of each generated image.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Generate Seedance 2.5 Image Video
>
> **Slug:** `HIGGSFIELD_GENERATE_SEEDANCE_2_5_IMAGE_VIDEO`
>
> Submit a ByteDance Seedance 2.5 image-to-video request through Higgsfield and return a request ID for tracking. Managed formula estimates use requested duration and a provisional reference-frame area and may differ from final supplier usage. Managed estimates above \$2.50 are rejected before generation submission. No polling is performed; use the returned request ID with GET\_REQUEST\_STATUS. Managed generation charges the upfront estimate when submission is accepted, including if generation later fails. Managed requests cannot be canceled.
>
> **Instant:** Supported on the latest version.
>
> **Instant pricing:** Seedance 2.5 image-to-video uses a numeric quote or a provisional 480p/720p reference-frame estimate at \$2.140 per 100,000 video tokens before discounts. Estimates above \$2.50 are rejected. Reference frame area may differ from the generated video.
>
> #### Input Parameters
>
> **prompt** `string`
>
> Optional description of the desired scene and motion.
>
> **duration** `integer`
>
> Requested output duration in seconds (4-30).
>
> **image\_url** `string` *required*
>
> Public URL of the starting image to animate.
>
> **resolution** `string ("480p" | "720p" | "1080p")`
>
> Output resolution tier.
>
> **end\_image\_url** `string`
>
> Optional public URL of the desired final frame.
>
> **output\_format** `string ("mp4" | "mov")`
>
> Video container format.
>
> **generate\_audio** `boolean`
>
> Whether to generate audio with the video.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Generate Seedance 2.5 Text Video
>
> **Slug:** `HIGGSFIELD_GENERATE_SEEDANCE_2_5_TEXT_VIDEO`
>
> Submit a ByteDance Seedance 2.5 text-to-video request through Higgsfield using the connected API key and return the request ID used to track the asynchronous result with GET\_REQUEST\_STATUS. Formula estimates use requested duration and estimated dimensions, and may differ from the final provider bill. Managed estimates above \$2.50 are rejected before generation submission. No polling is performed; use the returned request ID with GET\_REQUEST\_STATUS. Managed generation charges the upfront estimate when submission is accepted, including if generation later fails. Managed requests cannot be canceled.
>
> **Instant:** Supported on the latest version.
>
> **Instant pricing:** Seedance 2.5 uses a numeric quote or a 16:9 480p/720p upfront estimate at \$2.140 per 100,000 video tokens before discounts. Estimates above \$2.50 are rejected; other variants require a numeric quote.
>
> #### Input Parameters
>
> **prompt** `string` *required*
>
> Text description of the video to generate.
>
> **duration** `integer`
>
> Requested output duration in seconds (4-30).
>
> **resolution** `string ("480p" | "720p" | "1080p")`
>
> Output resolution tier.
>
> **aspect\_ratio** `string ("16:9" | "4:3" | "1:1" | "3:4" | "9:16" | "21:9")`
>
> Output width-to-height ratio.
>
> **output\_format** `string ("mp4" | "mov")`
>
> Video container format.
>
> **generate\_audio** `boolean`
>
> Whether to generate audio with the video.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Generate Seedance 2 Text Video
>
> **Slug:** `HIGGSFIELD_GENERATE_SEEDANCE_2_TEXT_VIDEO`
>
> Submit a ByteDance Seedance 2.0 text-to-video request through Higgsfield and return the request handle used to track the asynchronous result. Managed estimates above \$2 are rejected before generation submission. No polling is performed; use the returned request ID with GET\_REQUEST\_STATUS. Managed generation charges the upfront estimate when submission is accepted, including if generation later fails. Managed requests cannot be canceled.
>
> #### Input Parameters
>
> **prompt** `string` *required*
>
> Text description of the video to generate.
>
> **duration** `integer`
>
> Requested output duration in seconds (4-15).
>
> **resolution** `string ("480p" | "720p" | "1080p" | "4k")`
>
> Output resolution tier.
>
> **webhook\_url** `string`
>
> Optional public HTTPS URL that receives the terminal completed, failed, or nsfw payload.
>
> **aspect\_ratio** `string ("16:9" | "4:3" | "1:1" | "3:4" | "9:16" | "21:9")`
>
> Output width-to-height ratio.
>
> **generate\_audio** `boolean`
>
> Whether to generate audio with the video.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Generate Image
>
> **Slug:** `HIGGSFIELD_GENERATE_SOUL_STANDARD_IMAGE`
>
> Submit a text-to-image request to Higgsfield Soul Standard and return the request ID used to track the asynchronous result. Managed estimates above \$2 are rejected before generation submission. No polling is performed; use the returned request ID with GET\_REQUEST\_STATUS. Managed generation charges the upfront estimate when submission is accepted, including if generation later fails. Managed requests cannot be canceled.
>
> **Instant:** Supported on the latest version.
>
> **Instant pricing:** Soul Standard: varies by resolution and image count; \$2 max per request.
>
> #### Input Parameters
>
> **seed** `integer`
>
> Optional deterministic generation seed (1-1000000).
>
> **prompt** `string` *required*
>
> Text description of the image to generate.
>
> **style\_id** `string`
>
> Identifier of a style supported by Soul Standard.
>
> **batch\_size** `integer ("1" | "4")`
>
> Number of output images. Soul Standard supports only 1 or 4.
>
> **resolution** `string ("720p" | "1080p")`
>
> Output resolution. Use 720p or 1080p.
>
> **webhook\_url** `string`
>
> Optional public HTTPS URL that receives the terminal completed, failed, or nsfw payload.
>
> **aspect\_ratio** `string ("9:16" | "16:9" | "4:3" | "3:4" | "1:1" | "2:3" | "3:2")`
>
> Output width-to-height ratio.
>
> **enhance\_prompt** `boolean`
>
> Whether Higgsfield should enhance the prompt before generation.
>
> **style\_strength** `number`
>
> Strength of the selected style (0-1).
>
> **custom\_reference\_id** `string`
>
> Optional identifier of a completed character reference owned by the account.
>
> **image\_reference\_url** `string`
>
> Optional public image URL used to guide prompt enhancement.
>
> **custom\_reference\_strength** `number`
>
> Strength of the character reference (0-1). Zero uses the selected style's default character strength.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Generate Soul V2 Image
>
> **Slug:** `HIGGSFIELD_GENERATE_SOUL_V2_IMAGE`
>
> Submit a text-to-image request to Higgsfield Soul V2 and return the request ID used to track the asynchronous result. Managed estimates above \$2 are rejected before generation submission. No polling is performed; use the returned request ID with GET\_REQUEST\_STATUS. Managed generation charges the upfront estimate when submission is accepted, including if generation later fails. Managed requests cannot be canceled.
>
> **Instant:** Supported on the latest version.
>
> **Instant pricing:** Soul V2: varies by resolution and image count; \$2 max per request.
>
> #### Input Parameters
>
> **seed** `integer`
>
> Optional deterministic generation seed (1-1000000). Omit to use a random seed.
>
> **prompt** `string` *required*
>
> Text description of the image to generate.
>
> **style\_id** `string`
>
> Identifier of a style supported by Soul V2.
>
> **batch\_size** `integer ("1" | "4")`
>
> Number of output images. Soul V2 supports only 1 or 4.
>
> **resolution** `string ("720p" | "1080p")`
>
> Output resolution tier.
>
> **webhook\_url** `string`
>
> Optional public HTTPS URL that receives the terminal completed, failed, or nsfw payload.
>
> **aspect\_ratio** `string ("9:16" | "16:9" | "4:3" | "3:4" | "1:1" | "2:3" | "3:2")`
>
> Output width-to-height ratio.
>
> **enhance\_prompt** `boolean`
>
> Whether Higgsfield should enhance the prompt before generation.
>
> **style\_strength** `number`
>
> Strength of the selected style (0-1). This parameter currently has no effect on generation.
>
> **custom\_reference\_id** `string`
>
> Optional identifier of a completed character reference owned by the account.
>
> **custom\_reference\_strength** `number`
>
> Strength of the character reference (0-1). Must be greater than 0 when custom\_reference\_id is supplied.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Generate Video From Image
>
> **Slug:** `HIGGSFIELD_GENERATE_VIDEO_FROM_IMAGE`
>
> Generate a video from a publicly accessible starting image with Higgsfield DOP Lite and return the request ID used to track the asynchronous result. Managed estimates above \$2 are rejected before generation submission. No polling is performed; use the returned request ID with GET\_REQUEST\_STATUS. Managed generation charges the upfront estimate when submission is accepted, including if generation later fails. Managed requests cannot be canceled.
>
> **Instant:** Supported on the latest version.
>
> **Instant pricing:** DOP Lite video: up to \$2 per request.
>
> #### Input Parameters
>
> **seed** `integer`
>
> Optional deterministic generation seed (1-1000000).
>
> **prompt** `string` *required*
>
> Text description of the desired motion and scene.
>
> **image\_url** `string` *required*
>
> Public HTTP or HTTPS URL of the first-frame image to animate. The image must already be publicly accessible.
>
> **webhook\_url** `string`
>
> Optional public HTTPS URL that receives the terminal completed, failed, or nsfw payload.
>
> **end\_image\_url** `string`
>
> Optional public HTTP or HTTPS URL of the desired final frame. The image must already be publicly accessible.
>
> **enhance\_prompt** `boolean`
>
> Whether Higgsfield should enhance the prompt before generation.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Get Character Reference
>
> **Slug:** `HIGGSFIELD_GET_CHARACTER_REFERENCE`
>
> Get a Higgsfield Soul ID character reference and its current training status by ID using your connected account.
>
> #### Input Parameters
>
> **id** `string` *required*
>
> Character-reference UUID returned by CREATE\_CHARACTER\_REFERENCE.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Get Request Status
>
> **Slug:** `HIGGSFIELD_GET_REQUEST_STATUS`
>
> Get the current state of a Higgsfield generation request and return generated image, video, or audio URLs when complete.
>
> **Instant:** Supported on the latest version.
>
> **Instant pricing:** \$0 per request
>
> #### Input Parameters
>
> **request\_id** `string` *required*
>
> Generation request ID returned by a Higgsfield generation action.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] List SOUL V2 Styles
>
> **Slug:** `HIGGSFIELD_LIST_SOUL_V2_STYLES`
>
> List SOUL V2 image styles available to your connected Higgsfield account, including style IDs for SOUL V2 generation requests.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not
