đľ SongWriter â AI Song Creation with Full Creative Control (Alpha) Mimic PC Exclusive Workflow
âď¸ 1. Start with an idea â not a finished song
Everything begins with intent, not constraints:
- a simple song concept
- a theme, a character, or a story
- optional draft lyrics (or none at all)
No need to write a full prompt.
No need to think like a model.
You describe what the song is about â SongWriter handles the rest.
đ§ 2. Intelligent songwriting (lyrics + music prompt)
SongWriter uses advanced LLM reasoning to generate two distinct but perfectly aligned outputs:
- Professional-grade music prompt (ENGLISH only) Designed specifically for music generation models (Ace Step 1.5), with: structure instrumentation vocal characteristics mood and production intent
- Complete song lyrics Written in the language you choose: English, French, Spanish, Japanese, Korean, and more structured sections: Intro, Verses, Chorus, Bridge, Outro coherent storytelling, not random verses
Music intelligence and lyrical intelligence â separated, but synchronized.
đ 3. Language-aware by design
SongWriter understands that music models and lyrics models have different needs:
- đź Music prompts are always in English (for maximum model accuracy)
- âď¸ Lyrics are generated in the target language you select
- đ¤ The workflow outputs a 2-letter language code (en, fr, es, ja, etc.) ready to plug directly into Ace Step TextEncode nodes
No manual conversions.
No mismatched language settings.
đď¸ 4. Built for Ace Step 1.5 (not generic music AI)
Unlike âblack boxâ music generators, SongWriter is designed to integrate into a fully controllable synthesis pipeline:
- structured lyrics â Ace Step TextEncode
- clean music prompt â conditioning
- exact duration control
- seed control & reproducibility
- key, scale, and musical intent preserved
This is music generation as a system, not a lottery.
⥠5. Deterministic, editable, reusable
Every output is:
- visible (lyrics, prompt, full model output)
- editable at any step
- reproducible via seed
- compatible with downstream audio workflows
You are not âgenerating songsâ.
You are building songs.
đ§ 6. Final output: a real song, not a demo
At the end of the workflow you get:
- a fully generated song audio
- lyrics you can reuse, publish, or iterate on
- a music prompt you can tweak or reuse for variations
- a clean, production-ready result
No watermark.
No platform lock-in.
No creative ceiling.
đ Why SongWriter is different
- Separates music intelligence from lyrical intelligence
- Forces English-only music prompts for better audio results
- Supports multilingual lyrics without breaking the pipeline
- Designed specifically for Ace Step / ComfyUI, not generic AI music sites
- Full transparency â no hidden prompts, no magic steps
- One workflow, total creative control
This is not âAI music generationâ.
This is AI-assisted songwriting, engineered for creators.
Developed and optimized exclusively for Mimic PC.
