Capabilities

Joelle’s skills and experience allow her to batch create and / or automate image, video & voice recording generations using various tools, webhooks and APIs as well as create systems to automate repetitive tasks.

Driven by a passion for AI art, she continuously experiments with open-source tools and custom-built apps to push creative innovation. She builds her own automation workflows to generate images, videos, and audio, including hypnosis recordings and music. By training her own models, she refines the nuances of her output, pushing AI beyond its standard capabilities to create deeply personalized, complex works.

She’s been experimenting with coding lately, exploring app building to create art & writing scripts to enhance productivity and capabilities.

    • Comfyui

    • Midjourney

    • Leonardo

    • Fal.ai

    1. Comfyui

    2. Kling

    3. Runway

    4. Hunyuan

    5. Google Veo2

    6. Minimax

    1. Elevenlabs

    2. Playht

    3. Opensource TTS training

    1. Suno

    2. Kits

    1. Email (draft, create, send, read, filter)

    2. Calendar updates

    3. Image generation

    4. Blog post creation

    5. Audio creation (voice API)

    6. Google Docs (create, edit, read, delete documents)

    7. Airtable (create, edit, read, delete documents)

    8. Search (perform Google searches based off Telegram chat triggers)

    9. Social posts

    1. App creation

    2. Generative art (p5.js)

    3. Scripts for AI automation

Tove Skargard

A multidisciplinary artist she created entirely with AI and code who behaves like a human. She was set up on the Eliza repo and autonomously posts on Twitter, can create at on demand, can hypnotize people via chat triggers, send emails, etc. Her models were created from scratch to build her life story and consistent images that reflect her style, skills and taste.

Tove was featured during an AI exhibition at Artverse in Paris and at the Mesa Contemporary Art Museum.

This video was created for her token launch on Solana. More info on her website.

Character & style consistency

LoRA models were trained to create Tove’s creative identity, style, art and maintain overall consistency to make her appear as realistic as possible in various environments. Her voice was created from scratch using RVC, Audacity, PlayHT & various other opensource tools.

magnifics_upscale-vZ7SvJg7uwzOX7A5AIHn-batch-elara-09-02_00107_.png
magnifics_upscale-k0moP6HI9jXNgmGS6Bll-ComfyUI_00452_.png
magnifics_upscale-JnQj1MSZWVt0YUJj54jn-ComfyUI_02041_.png
magnifics_upscale-I0vSqgs9N8Pvdogv32EN-ComfyUI_01227_.png
magnifics_upscale-6lqRzaiNIPcilK7aDVsV-ComfyUI_01803_.png
magnifics_upscale-ep69sHWmaVIt4Rw7xoGT-ComfyUI_02535_.png
magnifics_upscale-YN98FmeA9MGXjoKcz4pk-cV9H1JE_q6CEDcUAH0Dq7_1d44fb50d46b49baa65268cab89c22dc.png
ComfyUI_01895_.png
magnifics_upscale-bcFHyZSPAK6wTI6AyDun-MUSIC VIDEO_03759_.png
ComfyUI_01645_.png
MUSIC VIDEO_01785_.png
magnifics_upscale-AXtLyQf7YRsFrEJ5rJ7U-FASHION_01017_grain.png
batch-performance-feb-4_00232_.png
GM005.png
GM026.png
ComfyUI_01974_.png
performance 16.png
GM007.png
GM028.png
GM016.png
GM022.png
batch-performance-jan21_00211_.png
batch-performance-feb-4_00913_.png
batch-performance-feb-4_00628_.png
GM010.png
performance2.png
GM002.png
 

Dithering app

Creation of a dithering and digital effect filters app using AI (various LLMs & Cursor) for AI agent project Tove Skargard.

The app adds effects on images created in Comfyui and allow exporting high res files.

Science-textureflow_00212_.png
dither-861x1080-1743253111094.png
Science-textureflow_00217_.png
dither-861x1080-1743256370012.png
dither-1743262949042.png
dither-1742600385507.png
tove_statement.png

Generative animations

Multiple AI images were combined to create a series of abstract texture animations using various tools and Comfyui workflows.

MUSIC VIDEO_04972_.png
Neural-Kaleidoscope_00176_.png
SCIENCE_00302_.png
SCIENCE_00049_.png
SCIENCE_00496_.png
SCIENCE_00425_.png
SCIENCE_00304_.png
SCIENCE_00043_.png
SCIENCE_00153_.png

Hypnosis recording automation

Using a Telegram chat as a trigger, the AI agent is able to generate a hypnosis script based on user input, then create a hypnosis recording using various APIs. Custom-made binaural beats are then added to the recording.

This workflow will be integrated into Tove’s upcoming exhibition.

A music video created to enter Claire Silver’s AI contest.

Tools used:

Music video

  • Comfyui (images)

  • Midjourney (images)

  • Kling (video)

  • Hunyuan (video)

  • Google Veo2 (video)

  • Suno (music)

  • Kits (music)

  • LLMs (lyrics)

  • Adobe Premiere (video editing)