I’ve built a lyrics tool specifically for Synthesizer V users who work with melodies.
What it does:
Write your own lyrics directly on a MIDI melody
→ The tool shows instantly whether each syllable fits the rhythm or not
Drop in text fragments
→ Place partial lyrics and check how they fit rhythmically – great for sketching ideas
Optional AI assistance
→ Describe what the text should be about, and the tool suggests lyrics that already match the rhythm of your melody. No AI needed though – works 100% manually too.
Direct output to Synthesizer V
→ Hear the result immediately, syllables already aligned
Technical info:
– Runs in the browser – platform independent (Windows, Mac, Linux)
– Communicates with SynthV via a script + shortcut (one-time setup, installer planned)
I’m looking for 5 beta testers willing to try it and give honest feedback.
Still in beta – but the core workflow already works really well.
Coolio – I totally see why you’d make that comparison! However, there’s a game-changing difference here: Hookpad can’t actually generate lyrics to fit an existing melody.
While Hookpad is a great notepad for chords and melodies, my tool works the other way around:
• Melody-Aware Generation: It ‘reads’ your finished melody and uses prosody-intelligence to ensure the lyrics don’t just match the syllable count, but also the rhythmic flow (matching fast notes with short words, etc.).
• Zero Manual Mapping: You don’t have to type or count a single syllable.
• The 14-Second Roundtrip: With the ‘M’ key script, the lyrics are injected directly into Synthesizer V, perfectly mapped to your phrasing.
Interesting idea, and why not?
Personally, when I compose, I always adapt my melodies to the lyrics and vice versa. This is to create something that isn’t dictated by the words.
I assume the link is located in a subfolder shared between the web service and the local disk.
Does the script read a file written by the web application? (Using Claude MCP).
It’s even simpler than a shared folder – no file I/O at all. The web app puts the lyrics directly into the clipboard in Synthesizer V’s native format. The Lua script then just triggers a paste via the assigned keyboard shortcut. That’s the whole “magic” – clipboard in, shortcut fires, lyrics land perfectly mapped in your project.
The setup is minimal: drop the script file into your SynthV scripts folder, then select it from the Scripts menu in the menu bar (or assign a hotkey to it in the Settings → Keyboard section – fully customizable). Hit “Rescan” once and it shows up. Done.
Regarding costs:
The manual workflow (write lyrics on MIDI, check rhythmic fit, inject into SynthV) is completely free.
For the AI-assisted suggestions, yes – there will be some kind of subscription, since managing individual API keys isn’t practical for the entry-level tier. The exact model is still being worked out. What I can say: during the beta phase, all AI calls are on me – beta testers get the full experience at no cost.
Ok, I was also thinking about this solution (and others… http etc.) when I was working on my own VST3 to interact directly with SynthV (and hoping that no other clipboard is used during the sequence…).
Yes, it’s as simple as that! I know a little about this subject (> 30 scripts in my repository).
Is another web server installed locally?
Ok, just wanted to know is a price is already known.
Thanks for sharing your view. I understand that AI in music can trigger strong feelings, especially for people who care deeply about songwriting and musicianship.
Just to clarify: this tool is not meant to replace songwriting talent or generate a finished song at the push of a button. The idea is much narrower.
It is designed for people who already have a melody and want help fitting lyrics rhythmically to that melody, especially in workflows like Synthesizer V. So the musical idea, the melody, the taste decisions and the final judgment still stay with the songwriter.
I see it more as a workflow assistant — closer to a rhyming dictionary, a syllable counter, a MIDI helper or an editing tool — not as a replacement for creativity.
I totally respect that some musicians don’t want to use AI at all. That is a valid position. For others, it may simply remove some technical friction while they remain fully responsible for the artistic result.
The difference is:
Instead of asking ChatGPT and then trying to force the text onto your melody, this starts from the melody and fits everything to it.
If you’re up for it, I’d really like to see how it performs on your actual workflow.
I’m currently keeping the beta very small (3–5 testers max), but if you’re using this kind of setup already, you’d be a perfect fit.
What’s new in the lyrics-on-MIDI tool (closed beta)
For anyone who followed this earlier: the project didn’t stand still.
The core idea is still the same — you already have a melody (e.g. in Synthesizer V), and the tool helps you get lyrics that actually fit the notes: syllables, timing, phrasing — with AI optional, not mandatory.
What got a lot better since spring:
Tighter melody fit — fewer broken syllable splits, better handling of SynthV quirks (melisma / -, compounds)
Generate → refine feels less chaotic — rhyme repair and retries are more predictable
Text stays intact more often after generate / repair / undo (less “why did that line get mangled?”)
Cleaner onboarding / demo path
You can leave in-app feedback; I actually see it
The landing page waitlist works properly again
It’s still closed beta — some parts are solid, some still rough. That’s honest.
I may open a few more tester spots again in the coming weeks for people who work with SynthV / MIDI on real projects. No rush — if you want to stay in the loop:
Write your own lyrics directly on a MIDI melody
Drop in text fragments
Optional AI assistance
Direct output to Synthesizer V