CANTO / SELF-INITIATED PRODUCT DESIGN
Advanced Suno AI Prompt Generator Working title for now
From a music idea to a prompt you can use.
I’m designing Canto to turn lyrics, feeling words, and musical intent into paired Styles and Lyrics prompts for Suno Custom Mode.
Approved flat baseline · Dark
I keep the Styles and Lyrics prompts beside the approved flat Mood Arc. This example keeps lyric mood and music mood in contrast.Approved flat baseline · Light
I designed the same Review experience in Light, including the prompt pair and approved flat Mood Arc.01
Why I’m building Canto
I want songwriters to be able to describe what they mean without first learning a producer’s vocabulary. Canto turns lyrics, feeling words, and musical intent into a pair of Styles and Lyrics prompts for manual use in Suno Custom Mode.
I’ve directed the work from the early Moodley identity through Canto’s brand, UX, and high-fidelity designs. I now have complete Dark and Light themes and a separate 3D Mood Arc prototype. I haven’t launched the product.
02
From an idea to a usable prompt pair
I start with the creator’s own words. Lyrics can arrive with section tags or without them, and I provide a way to check that structure without silently rewriting the song. I keep suggested lyric feelings separate from the person’s own keywords and their musical direction.
In Review, I put global Styles beside section-level Lyrics guidance, with a count and copy action for each. I treat a saved project as one song and History as its versions. Revisiting an earlier version should create another working version, rather than replace the source.
I provide a way to enter lyrics and check their section structure before shaping the prompt.
I separate Styles from Lyrics, with their own counts and manual copy actions for Suno Custom Mode.03
Keep the creator’s meaning intact
I framed Canto around feeling and intent because describing music in a prompt can become a vocabulary exercise. I want the interface to help people express the idea first, with more technical language in the resulting prompt.
I compared music and prompt tools and reviewed public-review material to help frame questions about inputs, output structure, and versioning. That was desk research, not a study with Canto users, so I don’t treat it as proof of demand or better results.
04
How I directed the work
I set the product direction, creative decisions, acceptance criteria, and approvals. I chose the wordmark’s geometry, spacing, and rounded direction, and kept earlier iterations available for comparison.
I used AI tools for assigned research synthesis, design execution, prototype work, exports, and checks. My job was to direct, evaluate, and decide. Ruby’s reviews helped me question the work and correct the documentation.
05
Keep the first version focused
I’m keeping the handoff simple: prepare the prompts in Canto, then copy them into Suno Custom Mode. I’m not including music generation, audio analysis, direct Suno integration, reference-track decoding, or a marketplace in this version.
I designed around budgets of 1,000 characters for Styles and 5,000 for Lyrics. I count the whole payload and keep the original when shortening. I use Medium and Tight treatments, with individual approval before removing lyric lines. I’m still working through the final Apply interaction.
06
From Moodley to Canto
I began with Moodley and later renamed the project Canto. I explored playful script and calligraphic lettering before moving to a more minimal geometric direction.
I asked for generous spacing and a consistent rhythm across the letters. When the A felt narrower than the others, I moved to custom equal-width construction. I chose the balanced direction, softened its edges, and kept the earlier versions so I could compare the decisions.
07
A wordmark and two complete themes
I refined the V021 wordmark around custom uppercase letters and a closed O based on the C’s circular geometry. I moved the O left to tighten its relationship with the T while leaving the other letters intact.
I paired the monochrome wordmark with Inter for the interface, spectrum-wave imagery, glass containers, and more opaque reading surfaces. I completed both Dark and Light themes and used warm primary actions to keep the next step visible.
I’m using “Advanced Suno AI Prompt Generator” as a working title for now. It is secondary to the Canto wordmark and isn’t a final tagline.
I refined V021’s closed O and the final T–O gap while keeping the custom uppercase construction.
I use a glass card over the Dark wave background, with a more opaque surface behind the reading area.
I carry the same glass-card treatment into Light while keeping the foreground readable.08
Keep each step clear
I separate creating, reviewing, saving, and revisiting. I let a guest start before asking them to sign in for permanent Save or private Saved and History access. I want sign-in to preserve their work and the action they were trying to take.
I kept blockframes, wireframes, and the design system in separate files so I could compare structure with visual decisions. I designed recovery, over-limit, account, and historical-version states across desktop and compact layouts. Authentication, storage, and deletion still need implementation validation.
I keep suggested lyric feelings separate from user-entered feeling words and show missing intensity sources as missing.
I let the musical direction deliberately contrast with the feeling in the lyrics.09
Show lyric mood and music mood separately
I use two strands because the words and the arrangement don’t have to express the same feeling. I place sections from left to right and use height for qualitative intensity, with a separate marker for tension. Labels and strand patterns carry meaning alongside color.
In my illustrative example, Verse 1 has sad, gentle lyrics with upbeat, intense music. The Chorus has hopeful, intense lyrics with calm, gentle music, with tension peaking there. Both sections stay in contrast. I’m showing a creative setting, not measuring audio or demonstrating emotional alignment.
I’m keeping the flat Mood Arc as my approved product baseline. The example is illustrative, not audio analysis.
10
Bring the design together
I use Phase 08 V002 as the product-design baseline. I brought the prompt pair, flat Mood Arc, saved songs, version History, and theme controls into a complete Dark and Light experience.
I checked readable exports and adjusted text sizes, dimming, and backing surfaces where the wave competed with content. I have a coherent screen set and design system to build from, with implementation and accessibility work still ahead.
I organize versions inside a song’s History. The storage behavior still needs to be built and validated.
I carry the same History structure into Light.11
Explore the Mood Arc in space and motion
I explored a separate 3D Mood Arc so I could look at the same lyric and music strands from different views. I use height for intensity and depth separation for their relationship, while keeping the underlying values intact.
I developed the experiment across Workbench and Review in both themes and four widths. In the prototype I explored playback, camera presets, and pointer and keyboard rotation, with reduced-motion behavior too. The static PNGs don’t make the depth or connector easy to judge, so I don’t present them as clear proof of the 3D interaction.
I’m keeping this experiment separate from the approved flat Mood Arc. I still need to improve the small projected axis labels and decide whether the 3D view belongs in the product.
I wrapped the compact Workbench controls into rows. This is the experimental 3D direction, and the small projected axis labels still need attention.Flat baseline and experimental 3D
Flat Mood Arc · approved product baseline
I keep the Styles and Lyrics prompts beside the approved flat Mood Arc. This example keeps lyric mood and music mood in contrast.3D Mood Arc · experimental visualization
I use a smooth strand for lyric mood and a banded strand for music mood. Height is intensity; depth is the relationship between them. This is a static crop of the experimental 3D design, so the depth is hard to judge here. The flat Mood Arc remains my approved product baseline.12
What I kept, changed, and set aside
I kept earlier directions so I could see how the work changed. Moodley’s playful treatment and Canto’s early calligraphy gave way to the geometric wordmark. I compared letter spacing, A construction, and edge treatment before settling on the rounded direction and V021 refinement.
I also moved through earlier Light-first and Dark-first interfaces before completing both V002 themes. Some versions capture execution problems, such as misaligned letters, misplaced archive content, or incorrect SVG transfer data. I keep those as part of the process, without presenting every saved attempt as an intentional design direction.
Rejected direction / Phase 07 Dark-First V001
I set this earlier dark-first direction aside before completing both V002 themes.
Rejected direction / Phase 08 Sound Wave V001
I set this version aside before the approved Phase 08 V002 direction. I’m keeping these rejected versions as text labels here, without images.
Pilot to rollout · experimental 3D
Historical approved pilot
I kept the approved earlier Workbench pilot so I could compare it with the later descriptor and control-layout changes.Approved experimental rollout
I carried the experimental Workbench direction into the full rollout, with the working title and controls filling their available rows.13
Where the work is now
I have an established wordmark, complete Dark and Light high-fidelity designs, responsive screen coverage, and documented product rules. I also have a separate working 3D visualization prototype.
I’m ready to use that work for implementation planning and further validation. I haven’t launched Canto or measured adoption, revenue, or usability gains. The flat Mood Arc remains my approved product baseline.
14
What I still need to work through
I still need to settle the signed-in starting point, sources for lyric and section intensity, version storage, deletion wording and retention, and the final tagline. I also need an approved example of alignment and reconnection, clearer continuous tension controls, and a final decision on lyric-cut Apply.
Before calling this a working product, I need to validate generation, sign-in, persistence, character counts, clipboard recovery, and responsive behavior. I’m keeping those questions visible as the next part of the work.