Higgsfield Launches AI Influencer Studio
On October 2, Higgsfield updated its AI Influencer feature, transforming it into a character studio for short videos. Characters can be built from scratch, based on reference images, or by uploading personal photos. An explanation on October 3 stated that uploaded faces will retain the facial features from the photo and require permission for use from the person in the photo.
Character creation does not require prompt words. The builder has 19 menu options, including character type, age, gender, skin color, eye color, face, body, hairstyle, accessories, and rendering style. Types are divided into three categories: Average, Bold, and Extreme, to control the intensity of unconventional features. Users can fill in randomly all at once. Each generation costs 1.12 points, approximately $0.06, and returns a close-up and a full-body image. As of October 3, each user has 5 free generations. Characters are stored in a character library and can also be accessed on a dedicated page in Cinema Studio.
Actions are completed on the Motion page within the same studio, based on Genjutsu. Action transfer extracts movements from reference videos, allowing characters to mimic them frame by frame. Object replacement swaps out people, products, or clothing in the video while leaving the rest of the footage intact. Output options include 480p, 720p, or 1080p. Reference videos can be between 4 to 30 seconds long and can include up to 30 images to describe characters, products, or clothing. There are also action templates available, eliminating the need for original filming. Voiceovers can be synced to lip movements.
This allows previously aired short video formats to be applied to new characters. The Explore page showcases finished characters and popular formats; clicking on one brings its settings back to the builder. Genjutsu was launched on September 1, stating that prompt words are not necessary, and users can choose from over 30 presets or simply describe the objects, locations, products, or characters to be modified. The advertising plugin simultaneously covers UGC, faceless videos, and bulk ad generation.
Character images and videos are created in two steps. Identical settings do not guarantee the same appearance in future generations. The help center states that both human and non-human characters can be created, with static images up to 4K resolution. The video side also features Motion Control, which combines character images with live performances, allowing characters to follow the actions while retaining their own appearance.
This is a paid creative tool, not a contracted influencer. Those paying are typically looking to produce short videos and product showcases in bulk, while Higgsfield collects payments based on points. The event is driven by characters, actions, and ad formats being consolidated on the same page. Beneficiaries include brands and agencies that require daily content without on-screen actors. Those under pressure are real creators who rely on UGC and influencers who charge based on facial consistency. The $0.06 buys a set of portraits, not a verified conversion.
Source: Public Information
ABAB AI Insight
Higgsfield initially launched Genjutsu on September 1, allowing users to swap people, products, and scenes in completed videos. On October 2, the character builder was integrated, replacing prompt words with 19 menus. The pathway is to first create reusable presets for actions and formats, then allow users to add a face. The 5 free generations serve as an entry point, while 1.12 points become the pricing unit thereafter.
Payments are made in points, not through influencer contracts. Each portrait costs about $0.06, with video pricing based on resolution. The motivation is to break down the two most expensive aspects of advertising—on-screen appearances and repeated filming—into menus and reference videos. Popular formats can be quickly filled in with settings, effectively leaving the structure of previously successful UGC for later use. Permission requirements are noted on uploaded photos, indicating that the source of the face is treated as a risk item by the product itself.
Comparable services include Captions and Arcads using synthetic voiceovers for ads, and Midjourney selling only static images while leaving motion to other tools. The difference is that Higgsfield consolidates building, action transfer, and object replacement within the same studio. Its industry position is during a period of production layer substitution. The actions of viral videos can now be transferred frame by frame, but the question remains of who is responsible for the authorization and advertising compliance of the face.
Structurally, this represents a reconstruction of the industry chain. Previously, a piece of UGC required someone to appear, someone to edit, and someone to distribute. Now, faces come from menus, actions from reference clips, and products from object replacements. The mechanism treats verified forms as templates, with new characters merely being skin swaps. The $0.06 reduces trial and error costs but does not alter the platform's traffic distribution. The remaining price difference for real influencers is their account history and the permitted identity, not the dance itself.
ABAB News · Cognitive Laws
Actions can be transferred, but account history cannot.
Once templates are clickable, on-screen appearances become a menu.
What is cheap is trying a face; what is expensive is who is responsible for that face.