Introduction

Generating a song with Suno or Udio can be fast and inspiring, but the difference between a messy result and something you can finish is how you write the prompt. Treat the prompt like a production brief: clear decisions where it matters, and deliberate gaps where you want the system to be creative. This guide gives a repeatable structure, practical decision rules, checklists and worked examples so you can write prompts that produce usable stems, arrangements and vocal tracks.

Why a structured prompt matters

When you start with an unspecific prompt, the model guesses everything at once—genre, tempo, instrumentation, arrangement and vocal phrasing. That makes outputs unpredictable and often unusable. A structured prompt reduces guesswork by assigning priority to the elements that are essential to your song. It also speeds up iteration because you know which part to change next: energy? remove or add a bridge? vocal style?

The five pillars of a music prompt

Focus your brief around five pillars. Address each one briefly in every prompt so the model has consistent constraints.

1. Genre and reference frame

• Be specific but not restrictive: name the genre (e.g., modern alt‑pop, neo‑soul, lo‑fi hip‑hop, indie folk), and optionally include 1–2 short artist references or tracks for style guidance.
• Use references as “style” rather than “copy”: say “in the style of Radiohead’s textural atmosphere” rather than “copy OK Computer”.

2. Energy, tempo and vibe

• Set the target energy: low, medium, high or numeric BPM range (e.g., 70–75 BPM for downtempo; 120–128 for dance).
• Define vibe with adjectives: intimate, aggressive, warm, cinematic, bouncy.

3. Arrangement and structure

• State the desired form: verse–pre–chorus–chorus–bridge, or loopable 2‑minute piece, or long evolving intro + 3 minute song.
• Indicate crucial parts: do you need intro, instrumental hooks, or a breakdown? If you plan to edit later, request stems or separate sections.

4. Vocal direction

• Decide: full lead vocal with lyrics, topline melody without words, or wordless vocal textures?
• Specify gender, range, timbre and delivery: e.g., female, mid‑range, breathy and intimate, with short, syncopated phrasing.

5. Sonic details and production cues

• Mention essential instruments or sounds (analog synth bass, flipacoustic guitar, warm tape delay) and inessential ones to avoid.
• Give a mixing cue—dry and intimate vs wide and reverbed—or leave mix decisions open if you plan to mix yourself.

Decision rules: what to specify and what to leave open

Not every detail needs to be locked. Use these rules to decide what to control and what to surrender.

1. Control anything you will edit later. If you need a precise vocal melody or drum groove to keep, specify it. If you plan to replace vocals anyway, leave the phrasing open.
2. Leave timbre and micro‑arrangement open for creative surprises. Let the generator produce interesting fills or pad textures you didn’t think of—unless they conflict with your concept.
3. Specify constraints for downstream tasks. If you plan to create stems, ask for separate stems. If you need a loopable instrumental, request a clean loop section.
4. Opt for ranges, not fixed numbers. Use tempo ranges, mood ranges (e.g., “medium‑low energy”) and soft descriptors rather than absolute statements.

Step‑by‑step workflow for drafting a prompt

Use this four‑stage workflow each time you prepare to generate. It keeps prompts focused and short while covering the pillars above.

1. Goal & constraints (30 seconds): One sentence stating the song goal and non‑negotiables. Example: “A 3:00 alt‑pop single with a clear chorus and lead vocal; no heavy EDM drops.”
2. Style & references (1–2 sentences): Genre and 1–2 references. Example: “Modern alt‑pop, moody electro textures; references: BANKS (vocals) and Lorde (minimal beats).”
3. Structure & parts (1–2 sentences): Form and what stems you need. Example: “Verse–chorus–verse–chorus–bridge; provide separate stems for drums, bass, keys, guitars and lead vocal.”
4. Tone & vocal direction (1 sentence): Energy, vocal delivery, and lyrical content if needed. Example: “Medium energy, intimate breathy lead vocal; lyrics about late‑night regret, 2nd person.”

Combine into a single prompt under 120–180 words. Shorter prompts are easier to iterate on.

Checklist: what to include before you press generate

• One‑line goal (song length and must‑have)
• Genre + 1 reference
• Energy/BPM range
• Structure and required stems
• Vocal direction and whether you want lyrics
• Two production cues (e.g., “warm, analogue bass”; “sparse percussion, lots of space”)
• One or two explicit avoidances (e.g., “no trap hi‑hats”)

Worked example: a pop single prompt

Here’s a real‑style example you can reuse and adapt. The goal is to create a prompt that gives enough control for a producer to work with the output without overconstraining the generator.

Goal: 3:15 contemporary pop single with a clear chorus and radio‑friendly vocal; provide separate stems for drums, bass, keys, guitars and lead vocal.
Style: Bright alt‑pop, in the vein of mid‑2010s Lorde meets Phoebe Bridgers—minimal beats, warm piano and textured synth pads.
Energy/BPM: Medium energy, 95–105 BPM, groove forward but not aggressive.
Structure: Intro (8 bars), verse (16), pre‑chorus (8), chorus (16), verse, pre‑chorus, chorus, bridge (8), chorus outro (16).
Vocal: Female lead, mid‑range, intimate breathy delivery with occasional close‑mic whispers; deliver full lyrics about a late‑night conversation and regret.
Production: Warm tape saturation on keys, tight dry snare, light plate reverb on vocals; avoid heavy EDM drops and trap hi‑hat rolls.
Output requirements: 24‑bit stems for drums, bass, keys, guitars, lead vocal and a mixed stereo version.

This prompt gives a short, ordered brief the system can follow while leaving arrangement details and fills flexible.

Worked example: an ambient interlude prompt

For ambient or experimental tracks you usually want looser direction. Here’s how to set intentional openness.

Goal: 2:30 evolving ambient interlude for a film cue; loopable 30‑second section in the middle.
Style: Warm, cinematic ambient with organic textures—soft field recordings, bowed glass, long pad swells.
Energy: Low, meditative, slow-moving; no percussion.
Structure: Slowly evolving intro build (0:00–0:45), loopable 30s center (0:45–1:15), further evolution to 2:30.
Vocal: Optional wordless alto vocal textures, mixed low and atmospheric—not a foreground lead.
Production: Lush low‑end removed below 60Hz for clarity, gentle stereo width; avoid sudden loud transients.
Output requirements: stereo mix and a 30s loop section as a separate file.

Note how essential constraints (length, loopable section, no percussion) are specified while timbral elements remain open to creative interpretation.

Iterative refinement and listening tests

Expect to iterate. Use a short test loop and a decision rule list to evaluate each output quickly.

• Listen for structure first: does the output contain the requested sections in order?
• Check the stems: are they clean and separated if you asked for them?
• Assess the vocal: is the delivery in the requested range and style?
• Confirm no prohibited elements (e.g., no EDM drops if you asked to avoid them).

If the result misses the mark, change one variable at a time. Example refinement sequence:

1. If energy is wrong, change BPM range or energy descriptor.
2. If mix is too wet/distant, add a production cue like “dry close‑mic vocals, minimal reverb.”
3. If arrangement lacks a hook, request “add a clear 4‑bar instrumental hook after the first chorus.”

Decision rules for lyrics, melody and performance

When to ask for lyrics vs. topline vs. wordless vocals:

• Ask for full lyrics if you want a final vocal you can plausibly release without rewriting.
• Ask for a topline melody (no words) if you plan to write custom lyrics or use the melody as a basis for human performance.
• Choose wordless textures if the voice should be treated as an instrument.

For melody specificity, use these rules:

1. Minor changes: if you want a slightly different chorus melody, request “same melody an octave higher” or “add countermelody.”
2. Major changes: if the melody is completely wrong, regenerate with a stronger reference (e.g., “Topline: catchy, stepwise melody with a rising phrase at the end of the chorus”).

Preparing generated material for production

Once you have usable stems or mixes, organize them for the next steps:

• Name files clearly (e.g., “lead_vocal_v1.wav”, “drums_stem_v1.wav”).
• Time‑align sections if the generator produced multiple versions; keep a master session file or marker list noting bars.
• If you plan to edit or master, consider moving stems into a workspace like Barichord Audio Lab for analysis, EQ and mastering work.

When you reach a releasable mix, you can use tools that automate promotion and content creation. For example, Barichord Autopost can turn a finished song into social clips, captions and YouTube SEO elements for scheduled release.

Common pitfalls and how to avoid them

• Over‑specifying everything: This can create flat, predictable outputs. Keep at least one area open for surprise (usually texture or fills).
• Too many references: Using more than two or three references pulls the prompt in multiple directions. Pick the dominant reference and one supporting reference.
• Ambiguous energy terms: Replace vague adjectives with examples or BPM ranges.
• Skipping output requirements: If you need stems, always say so. Assume mixed stereo output otherwise.

When to use a prompt tool or template

If you generate frequently, structured templates speed up iteration and keep results consistent. Tools that help craft prompts can guide you through the five pillars or export a compact prompt you can paste into the generator. For help building prompts with clear fields, visit the Suno Prompt Master to structure your brief before you generate.

Conclusion

Writing a better prompt is mostly about clear priorities: lock what matters, leave room for creative fills, and set concrete output requirements for what you will edit or release. Use the five pillars—genre, energy, arrangement, vocal direction and sonic cues—as your checklist each time. Start with short prompts, iterate with one change at a time, and organize outputs toward production. When you have a finished track, you can streamline promotion workflows with services like Barichord Autopost and clean up mixes in Barichord Audio Lab. Keep the brief concise, and treat each generation as a draft you can shape into a finished song.