How to Make a Production-Ready Song in Suno AI | Lyrical Literacy + Nana Coree Tutorial

Writing a song's lyrics is only half the job. Turning those lyrics into an actual playable, downloadable audio track is a separate step, and it's the one that trips people up if they don't know which fields to paste what into. This walkthrough covers exactly how to take a song generated in Lyrical Literacy, using the Nana Coree ghost artist modifier, and turn it into a finished production track inside Suno AI.

What you need before you start

Before opening Suno, three things need to come out of Lyrical Literacy: the clean lyrics, the performance tags generated by the meta command for your chosen artist, and the style string pulled from the session notes. For the example song in this tutorial, River Mumma Calling, a reggae jazz dub lullaby built with the Nana Coree modifier, all three pieces were already generated and ready to drop into Suno.

Where each piece goes in Suno

Inside Suno, clicking Create opens simple mode by default, but the workflow here needs advanced mode, which splits things into three separate fields: lyrics, style, and song title. The style string from the top of the session notes goes into the style of music field. For River Mumma Calling that string was reggae, dub, jazz, dub, spoken word, acoustic, steel pan. Suno reads this as a comma-separated list of keywords rather than full sentences, and each word shapes the sound, so it needs to stay under 120 characters and stick to keywords only. The full meta output, performance tags and all, goes into the lyrics field. Suno reads the tags inside square brackets as instructions: a tag like "vocal, behind the beat, intimate, unhurried" actually shapes the vocal performance, and a tag like "dub break, steel pan solo, bass" generates that instrumental section. The tags are doing the real production work, so they should never be stripped out.

Generating and evaluating the track

After selecting a vocal option, such as female, and adding a title, hitting create produces two versions at once. It's worth listening to both before deciding, since one version might land a better vocal take while the other nails a specific instrumental section like the dub break. Neither version is a final lock-in; more can be generated as needed. When evaluating a track, four things are worth checking specifically: whether the voice matches the intended tone (for example, a warm contralto, adjusted in the style field if it comes out too bright or thin), whether spoken-word sections were actually spoken rather than sung (fixed by wrapping those lines in parentheses inside the tag), whether a dub break or similar instrumental cue actually created its intended section (fixed by adding more instrument tags), and whether the outro faded the way the session notes intended (fixed by adding a fade-out instruction to the last line).

Fixing and finishing the track

If Suno cuts off the lyrics, the usual cause is that the total character count is too high, and trimming the longest section, often the spoken-word part, first tends to solve it. Once a version sounds right, there are three finishing options: Extend, which adds more content after the current ending if Suno cut the outro short; Get Stems, which splits the track into separate vocal and instrument files for remixing or releasing an instrumental version; and Download, which saves the finished audio as an MP3 or other audio file.

Key takeaways

• Three pieces move from Lyrical Literacy into Suno: clean lyrics, meta performance tags, and the style string from session notes.
• Suno's advanced mode has three separate fields for lyrics, style, and title; the style string goes into style and must stay under 120 characters of keywords.
• Bracketed performance tags inside the lyrics field are instructions that shape vocals and instrumental sections, not just labels.
• Suno generates two versions per request, and it's worth comparing both rather than settling for the first result.
• Extend, Get Stems, and Download are the three finishing tools once a version sounds right.

Try it yourself

If you've generated a song in Lyrical Literacy with a different ghost artist, such as Champa or Marley, the same workflow applies, just with a different sound driving the style and tags. Typing "list" into Lyrical Literacy shows every available ghost artist. This tutorial is part of the Lyrical Literacy series from Humanitarians AI.