Higgsfield AI emerged from stealth with an $8M seed round led by Menlo Ventures and launched Diffuse, a mobile character-driven video generator capable of inserting a user's likeness into stylized motion clips from a single selfie.
Higgsfield AI
Also known as: Higgsfield · Higgsfield DoP · Higgsfield Soul · Diffuse
Plain English
An AI video and image generation platform built around cinematic camera motion control—letting creators apply director-grade camera moves like Bullet Time, Dolly Zoom, FPV Drone, and SnorriCam to generated shots.
Technical Overview
Founded by former Snap generative AI lead Alex Mashrabov, Higgsfield trains spacetime Diffusion Transformer models conditioned on explicit 3D camera-trajectory and lens-physics embeddings (DoP I2V), alongside high-frequency texture and aesthetic photo models (Higgsfield Soul) and talking-avatar motion controls.
Timeline (4)
2024
Release- Release
Higgsfield released Diffuse 2.0, upgrading its foundational character-consistency and motion-transfer pipeline so creators could combine text prompts with reference photos and pre-choreographed motion templates.
2025
ReleaseHiggsfield launched its DoP (Director of Photography) image-to-video model (
DoP I2V-01-preview) on web, exposing over 50 named cinematic camera-motion presets—including Bullet Time, Crash Zoom, Dutch Angle, FPV Drone, and SnorriCam—as composable video generation controls.- Release
Higgsfield introduced Higgsfield Soul, a high-realism text-to-image model tuned with curated fashion, editorial, and candid smartphone photography presets that feed directly into Higgsfield DoP camera-motion video workflows.
See also (7)
Open full graphHiggsfield AI