Midjourney v7 Style Explorer, Seedream v4 Takes #1, ElevenLabs Video to Music
Nano Banana editing workflows, Hailuo keyframes, Zapier recognition, new portfolio showcase, and exclusive prompts inside.
Total reading time about 5 minutes.
Welcome to Visually AI!
Midjourney Style Explorer
Midjourney launched a Style Explorer for v7. It’s located on the Explore tab on the website and you can choose from ‘Random, Hot, Top’ style reference codes (—srefs) to try.
You can also choose ‘Randomize’ on the right side of the Explore page to get refreshed sets of styles to choose from or type a description in the Search bar, such as ‘cinematic, glitch’, etc.
Here’s a quick example of navigating Style Explorer:
These are all the same prompt with different sref codes I found when I typed in ‘editorial’ style in the search bar on the Style Explorer - srefs from top left and moving clockwise (2647866785, 153076371, 3562097221, 2224958557)
Seedream v4 takes the top spot on Image Editing leaderboards
Well, that was fast! Seedream v4 by ByteDance launched and rose to #1 on leaderboards, knocking Nano Banana down to second place (they seem to be taking turns at #1 at the time of writing this edition).
You can see the rankings here.
You can use Seedream v4 on most of the multi-model platforms (Replicate, Fal, Krea, Freepik, Dzine, Imagine, and more)
I don’t have any side by side comparisons yet, but Seedream appears to be quite capable of complex text renders. I tested it on Replicate with an image reference for the character and the prompt below.
Prompt: A panel of photos of this woman exploring a vintage vinyl record shop called “EchoTone Records.” A neon sign above the entrance reads “EchoTone Records - Live the Music.” Three panels: first, she’s outside the shop looking up at the glowing sign, then she’s inside flipping through vinyl sleeves where a wall poster says “Analog Forever,” and finally she’s seated by the shop’s listening station reading album notes from a record jacket. It’s all the same person and the record shop remains consistent across each panel.
ElevenLabs Video to Music | SFX v2
You can now easily turn a short clip into a cinematic scored video in minutes. Upload a ~10s clip to ElevenLabs Studio, select “Video to Music” and let it auto-generate a music track that matches motion and mood.
ElevenLabs released SFX v2 adds precise effects like whooshes, impacts, and UI taps so every cut feels intentional, not stock. Additional features include:
New Sound Library
Looping seamlessly
Max duration up to 30-sec
Export as MP3 or WAV on all plans
Nano Banana
I had some fun with Nano Banana editing combined with Kling’s Start & End frames, starting with my own photo:
Nano Banana (Gemini 2.5 Flash) — fast, on-model image editing
Google’s “Nano Banana” is a lightweight image-editing model built for speed and control. It handles background swaps, angle shifts, add/remove objects, and character continuity; so a face, outfit, or product stays consistent across scenes.
Why it matters
Continuity: keep the same character/style while changing context.
Speed: quick adjustments instead of full re-generations.
Control: object-level edits that don’t wash the whole frame.
Where you can use it:
Gemini app • Google AI Studio • FAL • Krea • Freepik • Replicate • Leonardo AI • Adobe Firefly • Higgsfield AI (per current rollouts), and many more platforms
I wrote about this in last week’s edition, for more details and links to each platform above.
Nano Banana is also available on Lovart AI. It’s simple and works well on Lovart’s ChatCanvas UI.
Select a region, hit tab, and type what you want. Let the agent do the rest...
Pro tips
Use short, concrete verbs: “add red cap,” “replace background with stadium,” “rotate 15°.”
Tight selections reduce bleed; iterate in small steps.
For batch edits, pair with Lovart ChatCanvas: select region → Tab → comment → Run All for consistent variations.
And here is a multi-element grid, 3-step workflow you can follow using Nano Banana to combine different elements in a single scene:
Step 1 - Grid theme prompt:
8-panel grid layout with clear grid lines separating each square:
Panel 1) sleek chrome samurai helmet centered on pure white background,
Panel 2) glowing neon katana centered on pure white background,
Panel 3) black armored chestplate centered on pure white background,
Panel 4) futuristic neon-lit Tokyo street at night,
Panel 5) cybernetic boots centered on pure white background,
Panel 6) robotic hawk centered on pure white background,
Panel 7) glowing holographic emblem centered on pure white background,
Panel 8) portrait of fierce cyber-samurai warrior centered on pure white background. Professional product photography style.
Step 2 - Grid to scene prompt:
Remove the grid lines and merge these elements into a cinematic cyberpunk scene. Show the samurai warrior wearing the full black armor with the glowing cyber-enhanced boots, wielding the cyan energy katana. Have the warrior releasing a cybernetic eagle companion that soars overhead with glowing wing trails. Position them standing on a rain-slicked street in the neon-lit Tokyo cityscape at night, with the holographic cube floating beside them as a data interface they're accessing. Create a dynamic action shot with the warrior mid-stride, rain droplets frozen in the air, neon reflections bouncing off the wet armor
Step 3 - Then I used a JSON prompt in Veo 3 to bring it all together:
This is another example of this workflow using different elements on 𝕏.
Visually AI named in Zapier's Best AI newsletters!
I want to say a big thank you to Zapier for featuring me in their top 10 AI newsletters of 2025!
Thank you all so very much for subscribing and reading Visually AI :)
You can see the full article with all the other top newsletters by clicking the button below:
🚀 My Recent Top AI Tools Picks
Hailuo Start/End Frames: Hailuo MiniMax dropped their keyframe feature with the option to generate with a start frame, end frame, or both. Here’s some of my first results, with audio added by ElevenLabs Video to Music:
Dia Browser: Runs on your desktop and lets you use AI to search, summarize, and organize info on any website you visit. I just started using this and it’s helpful so far. I have 5 invite codes if you want to try it free for yourself.
Envato AI Stack: Envato has full menu of generative AI tools including image/video/music/graphics generation, image editing, and more.
Mirelo SFX: Video to video audio generation syncs perfectly with the motion and tone of your video. Try the Mirelo app for free or via API on Fal.
I generated this audio with Mirelo on Fal:
🎨 Explore my Portfolio
I’ve had some amazing opportunities to work on a variety of projects recently, and I don’t share everything in this newsletter or online. And you know I love testing tools to share.
Take a look at my portfolio to get a quick glimpse of my work:
🖼️ Image Prompts
These grid panel images work well on image models with editing capabilities, such as Nano Banana and Qwen Image.
Prompt: 8-panel grid layout: Panel 1) golden silk gown on white background, Panel 2) ornate desert crown centered on white background, Panel 3) intricate golden arm cuffs centered on white background, Panel 4) endless desert dunes with glowing sunset, Panel 5) jeweled sandals centered on white background, Panel 6) majestic falcon with wings spread centered on white background, Panel 7) emerald pendant necklace centered on white background, Panel 8) regal woman’s portrait centered on white background. Each item distinct, professional photography.
Prompt: Combine all into a seamless desert fantasy portrait. Show the regal woman as the Desert Queen wearing the golden gown, crown, jeweled sandals, emerald pendant, and golden arm cuffs. A falcon soars above her. Place her atop desert dunes at glowing sunset, emphasizing mythic beauty and royal power.
Prompt: A lone figure sits in a high-tech lab, their face illuminated by the glow of a computer screen filled with complex code. The room is dark, save for the hum of machinery. Hacking thriller.
🎥 Video Prompts
Image to Video Prompt: Wide-angle shot: A breathtaking indoor waterfall cascades down from a glass ceiling, surrounded by lush greenery. The camera slowly pans upward, revealing the intricate geometric patterns of the structure above, while soft sunlight filters through, creating a gentle rainbow effect at the base of the falls.
Seedance 1.0 Pro
Image to Video Prompt: non-humanoid alien with crystalline pincer-digits gently holding a single white flower by the stem under a glass magnifier, sterile white lab bench, stainless tools, cool UV edge light and soft top fill, micro-condensation on glass, shallow depth, product-grade realism, disciplined negative space
Midjourney Video
Thank you for reading.
Have a creative week!
