CANTO / SELF-INITIATED PRODUCT DESIGN
Advanced Suno AI Prompt Generator Provisional descriptor
From branding to design-stage MVP. Canto is a design-stage AI music prompt translator: a way to turn lyrics, feeling words and musical intent into a paired Styles and Lyrics handoff for Suno Custom Mode.
Approved flat baseline · Dark
Review in Dark combines the prompt pair and flat Mood Arc. The two-section fixture is illustrative and remains in contrast; no alignment or reconnection is claimed. Approved flat baseline · Light
Review in Light combines the prompt pair and flat Mood Arc. The two-section fixture is illustrative and remains in contrast; no alignment or reconnection is claimed. 01 Giving songwriters a clearer way to direct AI music 02 Product overview: from intent to a usable prompt pair 03 Problem and opportunity: preserve the creator’s meaning 04 My role and the AI-assisted workflow 05 Product strategy and MVP boundaries 06 Naming and brand development 07 Visual identity: V021 within a complete two-theme system 08 UX architecture and interaction design 09 Mood Arc: making two kinds of intent visible 10 The high-fidelity experience 11 Experimental 3D Mood Arc 12 Design evolution: what changed and what was set aside 13 Current design-stage MVP status 14 Open decisions and next steps 01
Giving songwriters a clearer way to direct AI music Canto is a design-stage AI music prompt translator: a way to turn lyrics, feeling words and musical intent into a paired Styles and Lyrics handoff for Suno Custom Mode. I directed the product from its early Moodley identity through Canto’s brand system, UX architecture and high-fidelity experience. The work now includes complete Dark and Light themes, with a separate experimental 3D Mood Arc. It remains design and prototype work, not a launched product.
02
Product overview: from intent to a usable prompt pair The intended experience starts with a creator’s words. Lyrics can arrive with section tags or without them; the design provides a way to review structure without silently rewriting the song. Suggested lyric feelings remain distinct from the creator’s own keywords and from the musical direction. That distinction makes room for an upbeat arrangement beneath sad lyrics, rather than forcing both into the same emotional category.
The review screen separates global Styles from section-level guidance in Lyrics. Counts and copy actions sit with their respective fields. A project represents one song; its History contains versions, and revisiting an earlier version leads to another working version rather than replacing the source. These are designed product rules, not claims that production storage or generation has been implemented.
Lyrics input in Dark, with song-section structure visible. This is a designed input experience, not proof of implemented inference. Separate Styles and Lyrics outputs in Dark, with counts and manual copy actions for Suno Custom Mode. This does not demonstrate live generation. 03
Problem and opportunity: preserve the creator’s meaning My starting concern was the effort required to express musical intent in a prompt system’s vocabulary. The original feature record describes this as a language problem and sets “easier than Suno” as an ambition. It does not establish that the design achieved that benchmark. I used that ambition to frame the experience around feeling and intent, with expert terminology appearing in the resulting output rather than becoming a prerequisite for entry.
The project contains a dated competitive desk review spanning nine music-related tools and five general-purpose prompt tools. A companion document summarizes public-review material. These records helped frame questions about input methods, vocabulary, output structure and versioning; they are secondary research, not a study conducted with Canto users. The synthesis itself identifies missing review links, small samples and contradictory claims. I do not treat it as proof that competitors lack a capability, that a market gap is exclusive to Canto, or that creators prefer this design.
04
My role and the AI-assisted workflow I directed product strategy, creative decisions, acceptance criteria and approvals. My recorded feedback set the wordmark’s spacing and geometry, selected the rounded 02 direction, corrected the final O, and defined constraints such as manual copy-out, separate design files and preserved iterations. AI tools assisted with assigned research synthesis, design execution, prototype development, exports and verification. That division matters: directing and evaluating the work is different from personally drawing every vector or implementing every prototype.
Ruby’s reviews provided another check on the work and its documentation. Her latest corrections distinguish complete theme coverage from an unverified default, reopen the approval status of lyric-cut Apply, and require evidence for research claims. This case study incorporates those corrections rather than presenting the process as a sequence of uniformly final decisions.
05
Product strategy and MVP boundaries The design keeps the handoff bounded: Canto prepares prompts for manual use in Suno Custom Mode. Direct Suno integration, music generation, audio upload or analysis, reference-track decoding and a prompt marketplace are outside the documented MVP. The 440px extension surface is part of responsive design coverage; it is not evidence of a shipped browser extension or active automatic Fill implementation. Earlier extension-Fill planning was superseded by manual-copy guidance in the mid-fidelity handoff.
The project’s confirmed budgets are 1,000 characters for Styles and 5,000 for Lyrics. Counts include the complete payload, and over-limit output must not be silently truncated. Shortening preserves the original and offers Medium and Tight treatments where valid; removing lyric lines requires individual approval. The budgets were confirmed by Michael in the project record, not independently established as current platform limits by this documentation pass.
The final lyric-cut Apply interaction remains a proposal awaiting approval. Earlier files show an enabled button and use approval language, but Ruby’s latest correction expressly withdraws that interpretation as a settled UX decision. The reliable requirements are the character budget, original retention and explicit approval of cuts. The unresolved question is when Apply should become available, including whether selected cuts can be applied while the result remains over budget.
06
Naming and brand development The project began as Moodley and was renamed Canto. The branding explored playful script and later continuous calligraphic lettering before Michael requested a fresh minimalist direction. Those directions belong to the historical record; neither should lead the current identity story. The rename is documented, but a separate formal naming study, trademark clearance or naming-validation exercise is not established by the examined records.
The geometric wordmark developed through a more concrete sequence of decisions. Michael asked for generous spacing and a modern appearance, then pointed out that the A was narrower than the other letters. The response was custom equal-width letter construction rather than continuing to adjust the original font-based A. He selected 02 Balanced, then requested softened edges. The process log records approval of the rounded 02 direction on October 28 and its preservation in a separate production master. The rationale was the requested rhythm and geometry; no measured improvement in recognition is claimed.
07
Visual identity: V021 within a complete two-theme system The current V021 wordmark keeps the custom uppercase construction and uses a closed O based on the C’s circular geometry. After Michael noted that the O sat too far away, the refinement moved it left by four source units, changing the T–O gap from 58 to 54 while preserving C, A, N and T. Later transfer repairs corrected exported Bézier data without redesigning the Illustrator geometry. The verified-transfer files supersede the malformed transfer exports.
Canto’s approved V002 design includes complete Dark and Light themes. Inter carries the interface, while the monochrome wordmark remains visually separate from the spectrum wave imagery. Raster sound waves, glass containers and more opaque reading surfaces create the visual language. Warm primary actions are locked to amber #F3B65A with #17191D text in Dark, and orange #A74E08 with white text in Light. The sun/moon theme toggle has no checkmark. A locked default theme is not verified.
The Phase 09 rollout adds the provisional descriptor “Advanced Suno AI Prompt Generator.” It is secondary Inter text near V021. It is approved for that rollout for now, not confirmation of the final tagline or an assertion of direct Suno integration.
V021 light-surface geometry, with the closed circular O matched to the C and the refined T–O gap. A Dark glass card over the raster sound-wave background, with readable foreground content. This is a static design treatment, not runtime refraction. A Light glass card over the raster sound-wave background, with readable foreground content. This is a static design treatment, not runtime refraction. 08
UX architecture and interaction design The architecture distinguishes creating, reviewing, saving and revisiting. Guest creation comes before contextual authentication for permanent Save and private Saved/History access. The handoffs specify preserving the creator’s work and intended action through authentication, rather than restarting generation. Authentication providers, persistence and deletion services still require implementation validation.
Separate blockframe, wireframe and design-system files preserved the ability to compare structure and visual decisions. The mid-fidelity handoff records 16 screen categories across 85 states, including recovery, over-limit, account and historical-version situations. The high-fidelity record extends those states across 1440, 1024, 390 and 440px layouts. These counts describe design coverage; they do not represent 85 working application features or successful usability tasks.
The creator’s control over meaning remains central. Global Styles and local section instructions have different scopes. Inferred feelings are suggestions, not automatic replacements for the user’s words. The design also distinguishes a saved song from its individual versions, so revision can preserve context instead of overwriting the creative starting point.
Lyric feeling suggestions and user-entered feeling words remain separate in Light. Missing intensity sources are stated rather than fabricated. Independent music-mood direction in Dark; the arrangement can intentionally contrast with lyrical emotion. 09
Mood Arc: making two kinds of intent visible The Mood Arc was introduced as an explicit product requirement, not a conclusion from user research. Its two trajectories distinguish lyric mood from music mood. Section order runs horizontally; height represents qualitative intensity; tension is a separately labeled section marker. Direct words and strand patterns carry meaning alongside color. A crossing or visual proximity alone does not prove emotional agreement.
The approved progression fixture is deliberately limited. Verse 1 pairs Sad/Gentle lyrics with Upbeat/Intense music. Chorus pairs Hopeful/Intense lyrics with Calm/Gentle music, with tension peaking at Chorus. Both sections retain intentional contrast. These are illustrative creative settings, not audio measurements, and the example does not demonstrate alignment or reconnection. Missing lyric or section intensity must remain visible as missing information rather than being reconstructed from appearance.
Flat Mood Arc: approved product baseline. The fixture is illustrative, not audio analysis.
10
The high-fidelity experience Phase 08 V002 is the established product-design baseline for this narrative. The saved review records describe both themes across the four widths, with reading surfaces protecting lyrics, prompt outputs and counts from the wave behind them. The final export set includes the full prompt pair, flat Mood Arc, saved songs, version History, theme controls and visual specimens. The design-stage distinction remains explicit: native Figma glass is a visual treatment, not evidence of an implemented refraction engine.
Readable export inspection informed concrete corrections, including text sizing, dimming and opaque backing for labels. The historical report records representative contrast checks, but neither those checks nor attractive screenshots establish full accessibility certification. The outcome is a coherent, reviewable design system and screen set with documented implementation questions. It is not a measured improvement in conversion, task time or creative results.
Chronological versions within a song project in Dark. The version model is designed; immutable storage still needs validation. Chronological versions within a song project in Light. The version model is designed; immutable storage still needs validation. 11
Experimental 3D Mood Arc Phase 09 explores whether the same two mood strands can be understood in space and motion without changing their meaning. The V002 pilot introduced icon-over-label media controls, softer Dark depth, a clearer header hierarchy and a tension diamond. The reviewer qualifies the connector evidence: “connector reported in the local prototype / GPT report, not clearly visible in the static PNGs I visually inspected, where the diamond is close-anchored with its label.” The November 5 rollout brief records Michael’s pilot approval following Ruby’s visual review.
The full rollout applies that direction to Workbench and Review at all four widths in both themes. Michael’s additional Fill instruction distributes arc controls across their available rows, while text labels retain Hug sizing. The compact Workbench wraps six controls into three rows of two. Its measured header gaps are 32/48/32px at 1440 and 24/32/24px at the other widths.
Saved Chrome test records cover playback, next-action Play/Pause labeling, camera presets, pointer and keyboard rotation, reduced motion and preservation of the Chorus anchor and fixture. These are prototype checks, not creator usability results. The 16 rollout PNGs are available, but the last operational record shows publication of the separate rollout prototype blocked. The full Review experience is not a newly implemented application, and the 3D experiment has not replaced the production-design flat Mood Arc.
Experimental Phase 09 V002 Workbench at 390px in Dark, with the provisional descriptor and Fill control layout. The same two-section illustrative fixture remains in contrast; 3D has not replaced the flat product-design arc. Projected intensity-axis words remain a readability concern. Reviewer qualification: connector reported in the local prototype / GPT report, not clearly visible in the static PNGs I visually inspected, where the diamond is close-anchored with its label. Flat baseline and experimental 3D Flat Mood Arc · approved product baseline
Review in Dark combines the prompt pair and flat Mood Arc. The two-section fixture is illustrative and remains in contrast; no alignment or reconnection is claimed. 3D Mood Arc · experimental visualization
Experimental Phase 09 V002 Review at 1440px in Dark, with the provisional descriptor and Fill control layout. The same two-section illustrative fixture remains in contrast; 3D has not replaced the flat product-design arc. Reviewer qualification: connector reported in the local prototype / GPT report, not clearly visible in the static PNGs I visually inspected, where the diamond is close-anchored with its label. 12
Design evolution: what changed and what was set aside The archive preserves the decisions that led to the current work. Moodley’s bubbly treatment and Canto’s early calligraphy were retired when the brief changed. The calligraphy record also notes weak joins, overly mannered forms and pale hairlines; those observations describe the archived executions, not a general rejection of calligraphic branding. The geometric exploration then compared spacing, A construction and edge treatment before the rounded 02 selection and V021 refinement.
The interface similarly moved through historical Light-first and Dark-first directions before the complete V002 Dark and Light system. The earlier Gattaca-inspired palette exploration separated a restrained blue technology language from warm contrast accents; the later sound-wave and warm-CTA direction changed the visual treatment while retaining the product architecture. This history does not establish a current default theme.
Some retained versions document execution failures rather than creative alternatives: misaligned A construction, blank or misplaced archive content, and incorrect transfer SVG data. Their preservation makes the process inspectable without presenting every saved attempt as an intentional design milestone. Rejected and superseded work stays deeper in the story, clearly distinguished from approved identity and current product design.
Rejected exploration — Phase 07 Dark-First V001 Historical dark-first direction preceded the complete V002 Dark and Light system. It does not establish a current default theme.
Rejected exploration — Phase 08 Sound Wave V001 Set aside before the approved Phase 08 V002 direction. The requested comparison concerns glass treatment, background imagery and CTA hierarchy. The V002 package does not establish a more specific V001 rejection rationale; none is inferred here.
No images of either rejected direction are included.
Pilot to rollout · experimental 3D Historical approved pilot
Historical approved Dark Workbench pilot at 1440px, before the descriptor and Fill refresh. Retained as process evidence, not the latest rollout image. Approved experimental rollout
Experimental Phase 09 V002 Workbench at 1440px in Dark, with the provisional descriptor and Fill control layout. The same two-section illustrative fixture remains in contrast; 3D has not replaced the flat product-design arc. Reviewer qualification: connector reported in the local prototype / GPT report, not clearly visible in the static PNGs I visually inspected, where the diamond is close-anchored with its label. 13
Current design-stage MVP status Canto has an established wordmark, complete two-theme high-fidelity designs, responsive screen coverage, documented product rules, and a separately tested 3D visualization experiment. The approved flat Mood Arc remains the product-design baseline. The work is ready to support implementation planning and further validation, with its limitations visible. No production launch, adoption, revenue, user-study outcome or comparative usability improvement is claimed.
Phase 09 V002 Full Rollout was approved on Nov 5, 2026, about 09:34 ICT, by Michael after the reviewer’s visual review of all 16 rollout PNGs. Publication / hosted handoff is a separate decision and remains not authorized / blocked. This is not an approval conflict or a missing date.
14
Open decisions and next steps The next decisions concern behavior and evidence: the signed-in landing destination, lyric and per-section intensity sources, immutable snapshots, an approved alignment/reconnection fixture, continuous tension dragging, deletion wording and retention, the final tagline, advanced Suno examples and internal tag alignment. Final lyric-cut Apply approval and any locked theme-default behavior remain open. The provisional descriptor does not settle the tagline.
The 3D experiment also needs a separate production-adoption decision and further readability review, especially its small projected axis labels. Implementation and testing should establish the real behavior of generation, authentication, persistence, counts, clipboard recovery and responsive interactions. The case study documents what is designed and what has been tested in the prototype, leaving those next steps visible rather than converting them into claimed outcomes.
