WHISPERTECHAI / THE GENERATIVE STUDIO
Make something strange.
Make something useful.
AI isn’t interesting because a machine can generate a picture, a video, or a song. It’s interesting when an idea survives the trip from imagination to something you can actually see, hear, build, and share.
This is the creative side of WhisperTechAI: original experiments in generative video, character design, music, and working tools. Finished projects live alongside notes about what succeeded, what needed fixing, and what we’re still figuring out.
GENERATIVE FILM
CHARACTER STUDIES
AI MUSIC
REAL-WORLD BUILDS
FEATURED EXPERIMENT 001 / THE MUSIC LAB
Sizzle sa Kawali: When a cooking video got its own soundtrack.
The experiment started with a simple irritation: why keep pulling generic audio from an editing app when the actual footage could inspire its own music? We took a short cooking video from Kusina Cooks, developed an original funk-pop direction, generated a full song with Google Gemini, and worked its hook back into the cooking edit.
Sizzle sa Kawali · AI-generated cooking song on WhisperTechAI’s YouTube channel
ONE RECIPE. TWO DISTINCT RELEASES.
Real food. Generated music. Quite a lot of bass.
Oyster Beef with Kangkong provided the movement: garlic hitting the pan, the beef cooking, the greens folding into the sauce. The musical brief called for a punchy funk-pop rhythm at about 116 BPM, with a hook that could work inside a 19-second Short.
The full recording ran approximately two minutes. The cooking video took a matching excerpt, creating a connection between an actual meal and the AI music experiment.
Another song from the Kusina: We followed this up with “Init sa Palayok,” our AI-generated Filipino soundtrack for Sinaing na Tulingan. The new feature includes the music video, its creative story, and links back to the real cooking Short and recipe.
01 / SOURCE
Start with real footage
The video existed first. The food, cooking, and pacing gave the soundtrack a purpose beyond being background noise.
02 / DIRECT
Give the model a brief
Genre, tempo, instrumentation, vocal tone, and a hook designed for the first seconds of a Short became the creative constraints.
03 / PRODUCE
Make it fit the edit
We selected and matched a section to the cooking footage, while keeping the full song as a separate AI music release. The result is a workflow we can test again.
FEATURED EXPERIMENT 002 / FILIPINO AI MUSIC
Init sa Palayok: The recipe that became a song.
From real cooking footage to an original AI-generated soundtrack. A second experiment in matching music to the story of food, now presented alongside the process behind it.
Init sa Palayok · original AI-generated Filipino cooking song
ANOTHER SONG FROM THE KUSINA
A slow-cooked dish deserves a little fire.
The second collaboration started with Sinaing na Tulingan: pork fat and skin, dried kamias, fresh tulingan, and a patient simmer. Instead of selecting generic background audio, we created a Filipino-language AI song to give that cooking session its own musical identity.
The track embraces the spirit of the dish, though some lyrics are more poetic than procedural. That gap between a useful creative tool and a truly faithful representation is part of what we document here.
THE ARCHIVE / COLLECTIONS
More than a folder full of generations.
The private archive holds the experiments; the public Studio gives the strongest work an explanation. Every collection is curated around a question: what were we trying to make, what changed between versions, and what did the result teach us?
18-FRAME STORY + FILM NOW LIVE
Project Red / Red Moon
One of our earliest AI filmmaking projects: a red-maned wanderer, a shifting world, and a journey from isolation to conflict. The archive contains 18 keyframes across three acts and a finished 61-second vertical film.
Act I: Solitude and origin · Act II: Connection and discovery · Act III: The Red Moon and confrontation
The original eighteen frames are on display in their full three-act sequence, and the 61-second film is now available. Explore the Red Moon exhibit and watch the film, or watch the original Short on YouTube.
CHARACTER GALLERY NOW OPEN
Pinay AI / Character & Motion Lab
Filipina-inspired fictional characters across tropical portraits, food editorials, and futuristic identity studies. The gallery now features 13 original still images across three curated collections. We are preparing a separate Early Sora motion archive to study the videos without confusing them with the original stills.
The real question isn’t whether one portrait looks convincing. It’s whether a character, setting, and visual language can survive changes in angle, style, and motion. Explore the Pinay AI gallery.
Real or AI?
Visual perception tests and comparisons. What makes a scene convincing—and what gives it away?
Motion Studies
Our Early Sora archive is being curated: moving portraits, location and action studies, silent original generations, and edited reels. We will compare what worked with the awkward places where motion generation breaks, rather than simply publishing every file.
AI Nonsense
Absurd shorts, unexpected generations, and experiments that went sideways. Explore AI Nonsense.
HOW WE WORK
Prompts aren’t the whole story.
The interesting part begins after the first generation.
AI can speed up the first draft of an image, song, workflow, or application. It doesn’t automatically make the result coherent, accurate, original, or worth publishing.
- Curate, don’t dump. A hundred variants may belong in the private archive; only a few need to be on the public page.
- Show the decisions. We describe what we actually directed, edited, changed, or rejected instead of pretending a prompt did everything.
- Check the real limitations. Character consistency, uncanny details, music licensing, performance, and output quality deserve scrutiny.
- Keep building. A working app, a finished video, or a usable creative system tells us more than a list of promised capabilities.
For the more technical side of this work, explore our Builds & Thoughts section, including our guide to local LLM setups and experiments with persistent memory architecture.
AN EVOLVING CREATIVE LAB
More experiments are on their way.
Project Red is now open, with all eighteen original frames to explore. Pinay AI, visual comparisons, and new creator tools will join as their media and editorial notes are reviewed. We would rather publish a small collection with a story than a thousand thumbnails without one.
Creative disclosure: Some images, videos, voices and music featured in the Studio are generated or transformed using AI tools. Human creative direction, editing, selection and documentation shape the published work. Specific generation tools are identified when confirmed; media ownership and publication rights are reviewed before public exhibits are expanded.