Higgsfield: Generating AI video with tagged character references

Higgsfield Try it →Watch full video →Creative Pad Media · Mar 2026

Demo summary

The creator shows how to paste a multi-shot prompt into Higgsfield and use the '@' symbol to tag specific character images (detective and suspect) before generating the final video.

Step-by-step

Paste your multi-shot prompt into the first text slot
Locate the character names within the prompt
Tag the first character image by typing '@' and selecting image one next to the character name
Tag the second character image by selecting image two next to the corresponding character name
Select the desired resolution and aspect ratio
Click Generate

Options

Adjust resolution settings
Change the aspect ratio

Tips

You only need to tag the character image once within the prompt for the AI to recognize it

Highlights

“You're pretty much done. And you hit generate.”

All demos from “Cinematic Conversation Scenes Using Kling 3.0 Omni - Quick Method”

0:481:24Setting up multi-shot video in Higgsfield AIThe user demonstrates how to select the Kling 3.0 Omni model in Higgsfield, upload character reference images as 'elements', and enable the multi-shot feature for scene generation.Higgsfield· AI Animation Generator
3:360:49Generating AI video with tagged character referencesCurrentThe creator shows how to paste a multi-shot prompt into Higgsfield and use the '@' symbol to tag specific character images (detective and suspect) before generating the final video.Higgsfield· Text to Video
6:570:33Creating character assets with voice in Kling AIThe demonstration shows the native Kling AI 'Principal Assets' section where a user can upload a character image and a custom voice clone from ElevenLabs to ensure audio-visual consistency.Kling AI· AI Avatar Video Generator
7:301:17Multi-shot prompting in Kling AI native appThe user demonstrates pasting a full multi-shot prompt into the Kling AI native interface, tagging a saved character asset, and adjusting aspect ratio settings.Kling AI· Text to Video
Watch “Cinematic Conversation Scenes Using Kling 3.0 Omni - Quick Method” →

Text to Video

2:50:294:47Generating emotional character performancesThe user demonstrates prompting Higgsfield AI for specific emotional physical cues, such as a trembling lip and sobbing, to create a dramatic character breakdown.Higgsfield AI
2:000:41Text-to-video with multi-shot in Kling 3.0The user demonstrates generating a dynamic military action scene using a text prompt and the new multi-shot feature in Kling 3.0 to create automatic scene cuts.Higgsfield AI
5:050:49Complex camera movement controlThe user prompts a robotic arm scene with specific lateral passes, crash pushes, and pull-backs to demonstrate Kling 3.0's improved camera physics and motion smoothness.Higgsfield AI
2:160:42Text-to-Video with SeaDance 1.5 ProThe video shows how to configure settings like aspect ratio, clip length, and audio in SeaDance 1.5 Pro to generate a video from a text prompt.Malva AI
5:110:30Generate cinematic locations with HiggsfieldThe creator shows how to use the Location option in Higgsfield Cinema Studio to generate a cinematic background with specific lighting and camera angles based on a short text prompt.Youri van Hofwegen
3:360:49Generating AI video with tagged character referencesCurrentThe creator shows how to paste a multi-shot prompt into Higgsfield and use the '@' symbol to tag specific character images (detective and suspect) before generating the final video.Creative Pad Media
1:260:24Generate subtle human emotion with Kling 3.0Kling 3.0 is shown performing a text-to-video generation of a young woman shifting from joy to concern, demonstrating realistic facial micro-expressions.Joseph Martin
8:344:13Generate an atmospheric establishing shot with Seedance 2.0A live demonstration of generating a 15-second video of a 'Storm Manor' using a text prompt, showing the real-time wait and the final rendered output featuring lightning and darkness.JSFILMZ