G Glenn Williams
Work Characters Games Work with me Contact
The bench

Toolkit

This is what I actually use, and everything I have access to. What each tool is good at, and where it fights me. My brand partnerships are disclosed right here on the page, so you know exactly where I stand.

most used

Where I'm partnered

Partnered
  • AdobeFirefly Ambassador
  • Topaz LabsCreative Partner
  • RunwayMLCPP
  • Nim VideoCPP
  • DreaminaCPP
  • Flova AICPP
Not partnered, paid by choice
  • ElevenLabs
  • Krea
  • Suno
  • CapCut

The unpartnered tools are on this page because they earn it, nobody is paying for those entries.

Image

Krea 2 Large

My image model for the Creature Comforts and Old Magic series. Krea reads tokens more like an artist than a literal transcriber, and paired with moodboards it builds art styles you can actually call your own. Character consistency is where it fights you, but you can train a LoRA to hold a face in place.

Nano Banana 2

The balance model, legible detail that survives hard lighting, and my character consistency and reference workhorse. I use it to build keyframes that hold style and character across a video, hook images for articles and posts, and the scenes for my Hidden Object puzzles, since you can edit an image without it changing everything else you already had right.

Adobe Firefly / Boards

Home base. Boards is a massive canvas where I generate images and video side by side and keep every reference laid out in front of me, and because Firefly runs inside the rest of the Adobe suite the whole toolchain connects without exporting between apps. I build the bulk of my social images here, usually Nano Banana 2 on the Boards canvas. The content filter is stricter than most platforms, but that is the tradeoff for being one of the few tools that can call itself commercially safe in most cases, so now and then a flaggable token stops a prompt, and it is worth it.

Video

Seedance 2.0

Where my videos get made, run daily through Runway. Multi-shot prompts with hard cuts, full shot lists, audio direction, and it holds up on complex scenes that break other models. Its favorite failure is spawning a second copy of your character, so constraint language earns its place. People will tell you Seedance ignores negative prompts. It has no separate negative field, but it does respect what you tell it to leave out, written into the prompt, and I use that every day.

Audio

ElevenLabs v3

My voice tool. I use it for voiceover on my Stor-AI Time videos and brand work when a piece needs narration, and I reach for its sound effects too. v3 is the expressive one, worth the extra prompt effort when the read has to carry feeling.

Suno

My music tool, for soundtracks and full tracks behind my music videos. I have my own thing I keep coming back to, a kind of acid jazz lofi, and v5.5 is the model that gets me there. Outputs can still need a final mix, but more often than not the work is minimal.

Upscaling

Topaz Astra 2 / Starlight Precise 2.5

My finishing pass, and I lean on both equally. Astra 2 is the creative upscaler, you can prompt it to guide the generation, so it does more than clean up, it makes decisions. Starlight Precise 2.5 runs purely off the input and upscales incredibly, no prompting, just faithful. I run the standalone Topaz app for professional work, but Topaz shows up inside so many of the platforms I already work in that I can upscale wherever I am without breaking stride, at quality that holds. It can take its time to render, but the output earns the wait.

Editing

Adobe Premiere Pro

My main editor on the PC. Where the shots get assembled into something that holds together.

CapCut

The mobile counterpart to Premiere. When I'm editing away from the desk, this is where it happens.

Claude

My powerhouse assistant. This whole site was built by describing it, and when a project needs a tool that doesn't exist, a puzzle zone editor, an answer-key collector, a form to log how each video got made, I build it here. The work behind the work runs through this.

The full map

The curated list is what I reach for. This is everything on the bench, captured from the actual model menus on each platform.

Adobe Firefly
Image
  • Firefly Image 5, plus earlier Firefly Image versions
  • FLUX.2 pro and FLUX.1 Kontext, plus earlier FLUX versions
  • Gemini 3.1 with Nano Banana 2, plus earlier Gemini and Nano Banana versions
  • GPT Image 2, plus earlier versions
  • Runway Gen-4 Image
Video
  • Firefly Video
  • Kling 3.0 line
  • Luma Ray 3.14, plus earlier Ray
  • Runway Aleph 2.0 and Gen-4.5, plus earlier
  • Gemini Omni Flash
  • Veo 3.1 / 3.1 Fast
Audio
  • Firefly Speech (Adobe, commercially safe)
  • ElevenLabs Multilingual v2 (provider: ElevenLabs)
  • Generate soundtrack (beta)
  • Generate sound effects
  • Translate (audio and video localization)
  • Adobe Podcast Enhance v2, linked from Firefly
Upscaling
  • Firefly Upscaler (Adobe)
  • Topaz Bloom
  • Topaz Gigapixel
Editing
  • Timeline, the built-in video editor
RunwayML
Image
  • Nano Banana 2 family (provider: Google)
  • GPT Image 2, plus earlier versions
  • Grok Imagine
  • Gen-4 native line
  • FLUX.2 Max / Klein
  • Seedream 5.0 Pro / 5.0
Video
  • Gen-4.5 native line, plus Characters, Act-Two, and Aleph 2.0
  • Seedance 2.0 / Mini / Fast
  • Kling O3 and 3.0 lines, plus earlier versions
  • Grok Imagine 1.5
  • Gemini Omni Flash
  • Veo 3.1 / 3
  • WAN 2.6 / 2.6 Flash
  • HappyHorse 1.0
Audio
  • Seed Audio 1.0 (ByteDance)
  • Lyria 3 Pro and Lyria 3 Clip (Google)
  • ElevenLabs Multilingual v2
  • ElevenLabs Text to SFX
Upscaling
  • Topaz Upscale to 4K
  • Magnific (Precision v2, Creative, Style Transfer)
Editing
  • Runway Studio
Nim Video
Image
  • Nano Banana 2
  • GPT Image 2 Medium
  • Seedream 5 Lite
  • Z-Image Turbo
  • Wan 2.7 Pro
  • Flux 2 Pro
  • Grok
  • Imagen 4 Ultra
  • Qwen 2 Pro
  • Higgsfield Soul High
  • Runway High
  • Hidream High
Video
  • Veo 3.1 Sound
  • Seedance 2.0 Standard
  • Kling 3 line, plus earlier versions, with keyframe and reference controls
  • Gemini Omni
  • HappyHorse 1.1
  • Wan 2.7 Std
  • Grok
  • LTX 2 Sound
  • Hailuo 2.3 Pro
  • Pika
  • Runway
  • Vidu Auto
  • Luma Ray
  • Hunyuan Anime
Lip sync
  • HeyGen
  • VEED Fabric
  • PixVerse
Upscaling
  • Topaz (premium, best quality)
  • Bytedance
  • Standard
ElevenLabs
Image
  • GPT Image 2, plus earlier versions
  • Nano Banana 2 family
  • Krea 2 Large / Medium
  • FLUX.2 Pro and FLUX.1 Kontext Pro
  • Runway Gen-4 Image / Turbo
  • Seedream, Kling, and Wan image models listed as Restricted
Video
  • Gemini Omni Flash
  • Veo 3.1 line
  • Runway Gen-4.5 / Gen-4 Turbo / Aleph / Act-Two
  • LTX 2 line, including Audio-to-Video
  • Seedance, Kling, and Wan lines all listed as Restricted
Audio
  • Text to Speech: Eleven v3, Multilingual v2, Flash v2.5, Turbo v2.5, plus earlier v1 and v2 models
  • Music v2, plus Music v1
  • Sound Effects
  • Voice Isolator
  • Voice Changer
Lip sync
  • Creatify Aurora
  • HeyGen Avatar 4
  • Sync 3
  • Veed Fabric
  • Sync Lipsync 2 Pro
  • Veed Lipsync
  • OmniHuman 1.5 (Restricted)
Editing and upscaling
  • Topaz Video Upscale (4K)
  • Topaz Image Upscale (4K)
  • Background Removal
  • ElevenLabs Studio
Dreamina
Image
  • Image 5.0 Lite, plus earlier Image 4.x versions
Video
  • Seedance 2.0 / Fast / Mini
  • Seedance 1.5 Pro with synced sound, plus earlier versions
  • Omni reference control

Seedance's native home, unrestricted.

Flova AI
Image
  • GPT Image 2
  • Nano Banana 2 / Pro
  • Seedream 4.5
  • Midjourney V7, the only Midjourney access on the map
Video
  • Seedance 2.0 line (premium)
  • HappyHorse 1.1
  • Kling 3.0 Audio / Omni
  • Vidu Q3 line
  • Grok Imagine Video
  • Veo 3.1 line
  • Gemini Omni Flash
  • Wan 2.6
  • Hailuo 2.3
  • OmniHuman 1.5 lip sync
Music
  • Suno
  • Mureka
Audio
  • Seed-Audio 1.0
  • ElevenLabs
  • MiniMax Speech 2.8 HD
  • Doubao
Krea
Image
  • Krea 2 Turbo / Large / Medium, plus Krea 1
  • Nano Banana 2 family
  • Imagen 4 line, plus earlier versions
  • Flux 2 line, plus earlier versions
  • Seedream 5 Pro / Lite / 4.5
  • Luma UNI-1
  • Ideogram 4.0
  • Recraft V4
  • Qwen 2512
  • Z Image
  • Kling O1
  • Multiple Models toggle, one prompt across several models at once
Video
  • Krea Realtime, real-time generation, exclusive to Krea
  • Tiers spanning Seedance 2.0 line, Kling 3.0 / O3 lines, Runway Gen-4.5, Veo 3.1 line, Hailuo 2.3, Wan 2.6 / 2.5, Vidu Q2 / Q3, LTX-2 line, and Grok Imagine, plus a legacy tier of earlier versions
Upscaling: generative enhancers
  • Krea Enhance, Krea Legacy, Krea Video (Krea native)
  • Magnific and Magnific Precise (Freepik)
  • Topaz Generative, Topaz Bloom
  • Topaz Starlight / Starlight 2.5
  • Topaz Astra / Astra 2
Upscaling: simple upscalers
  • Upscale V1
  • Topaz (to 22K)
  • Topaz Video
  • SeedVR2 (ByteDance)
Suno (standalone)
Music
  • v5.5
  • v5
  • v4.5+
  • v4.5-all (best free)
  • Plus earlier versions
Topaz (standalone)
Apps
  • Topaz Video, Topaz Photo, Topaz Gigapixel, Bloom, Mosaic, Topaz Image Web, Gigapixel iOS
  • Astra, with models Astra 2 and Astra 1, running Precise and Creative modes
  • Starlight model family in Topaz video tooling: Precise 2.5, Fast 2, Proteus, plus earlier versions
  • Image app workflows: Wonder 3, Super Focus v3, Face Recovery v3, Denoise, Relight, Colorizer, Bloom Creativity, Background Remover
CapCut
Editing
  • Mobile editor (ByteDance)

Six ways to Seedance

  • Dreamina (native, unrestricted)
  • Runway (open, my daily route)
  • Nim (open)
  • Krea (open)
  • Flova (premium)
  • ElevenLabs (restricted)

Topaz is everywhere I work

The standalone suite, plus embedded in Firefly, ElevenLabs, Runway, Nim, and Krea.

Worth knowing

  • Krea Realtime is Krea-only
  • Midjourney V7 access is through Flova only
  • OmniHuman 1.5 lip sync is on Flova and ElevenLabs
  • Nano Banana 2, GPT Image 2, Seedance 2.0, Kling 3.x, and Veo 3.1 show up almost everywhere

See what this stack makes. And if you are a brand wondering what it could make for you, start at work with me.