The problem
AI-music work was fragmented across prompts, notes, generators and disconnected versions.
Kordra Music
LiveAn AI-assisted planning and pre-production workspace for developing lyrics, song structure and creative direction before the audio is generated.
- My role
- Founder and Product Lead
- Platform
- Web application
- Built with
- Next.js, Supabase, Stripe, Vercel
My approach
Build a structured workspace around the song, keeping planning, review, variations and handoff connected.
The result
A live product with 900+ songwriters registered as of 3 August 2026.
The decisions
behind the product.
AI music generators can create an impressive first result quickly, but the work around that result often becomes fragmented across prompts, notes and disconnected versions.
Build a workspace, not another generator
Kordra Music concentrates on the work before and after generation, where continuity and judgement matter most.
Keep song form visible
Sections, production cues and lyrics remain close together so the draft reads like a song in progress.
Make feedback explain itself
Review signals sit beside structure and energy evidence rather than presenting an unexplained score.
Protect the working draft
Variations can be explored and saved without forcing each experiment into the main version.
From thinking to making.
A guided Blueprint before generation
Song intent, style and voice decisions are made explicit before work moves into the editor or an external generator.
Structured writing and review
A section-aware editor, coaching and review tools create a coherent revision loop.
Reversible creative branches
Alternative directions can be applied or saved as branches without erasing the working song.
More implementation detail
Portable handoff
Export options carry the useful song context into the next generator, document or studio step.
What came out of it.
Live product
Kordra Music is available at kordra.io and continues to evolve around the structured work between idea and handoff.
Registered users
900+ songwriters on Kordra Music, recorded 3 August 2026. This is not an active-use, revenue or growth figure.
A distinct product promise
Separated the original prompt-planning utility, Suno Architect, from Kordra Music’s broader workflow for planning, review, variations and handoff.
Connected creator workflow
Brought structured writing, review, variations and portable export around one working song rather than disconnected prompts and notes.
Constraints, lessons and next steps
The constraints
- Founder-led resources required a narrow product promise and careful prioritisation.
- Model cost, user trust and variable AI output made human oversight essential.
- Exports had to preserve portability rather than lock work inside Kordra Music.
What I learnt
- The first proposition focused on producing better prompts for Suno.
- Practical use shifted the centre of gravity towards the song itself: its structure, revisions and handoff.
- The product and company identity were separated so Godsall.dev could explain Adam’s decisions while Kordra Music serves product customers.
- Workflow-level problem framing rather than feature chasing.
- Positioning and product-boundary decisions grounded in user control.
- Founder-led delivery that carries strategy into an evolving live product.