Describe
One line of intent. Optional artist reference, optional locked lyrics, optional genre — or nothing at all, and let the pipeline pick.
The AI music pre-production engine
HookGenius is the pre-production engine for AI music. It writes Suno-ready lyrics, style prompts, song titles, structural metatags, and vocal direction in a single Claude Sonnet pipeline — refined across thousands of producer sessions. You paste the output into Suno, Udio, or Riffusion and press Create. No more ChatGPT round-trips. No more guessing which tags trigger the sound you actually want.
No credit card · 5 free generations · 30 seconds to your first Suno paste
See the 3 stepsThe truth, in three steps
HookGenius is the step before Suno — the part that decides whether the track sounds like a producer made it, or a prompt-typer did.
One line of intent. Optional artist reference, optional locked lyrics, optional genre — or nothing at all, and let the pipeline pick.
HookGenius runs a multi-stage Claude Sonnet pipeline that writes the lyrics, style prompt, title, structural metatags, and vocal direction as one cohesive vision.
Copy the lyrics block and the style prompt into Suno, Udio, or Riffusion. Press Create. The audio model handles the render.
HookGenius does not produce audio files. Suno does. Udio does. We make the part that decides how good the audio comes out.
Versus iterating with a generic LLM
ChatGPT will give you lyrics. It will not give you the six things below in one pass, formatted correctly, every time.
A dedicated producer pass enforces gender, texture, delivery, and register together — the triple-stack rule — so an Olivia Dean reference does not come back as country-rock.
Seventy-plus genre profiles, each tuned for the specific Suno tags that actually trigger the sound — not the tags that sound like they should.
[Intro], [Verse], [Pre-Chorus], [Chorus], [Bridge], [Drop], [Outro] placed for the song’s BPM and energy arc, with pipe syntax for instrumental sections and spacer dot lines that hold the space Suno would otherwise strip.
A banned-list at the tonal layer catches “rise from the ashes”, “neon lights”, “ghost in the machine”, and the rest of the AI-music sediment before it ever reaches your lyrics.
A separate album-architect stage holds genre, vocal identity, and sonic world across every track — so a ten-song project sounds like one artist, not ten.
No [Spoken] tags on anthemic choruses. No bracket descriptions sung as lyrics. No duplicate pronunciation tags that bleed across languages. The formatting is the music.
Style: melodic rap, hip hop, R&B emotional, atmospheric [Verse 1] Walking through the city neon lights Rising from the ashes of my past tonight Ghost in the machine I’m finally free Every shadow that I cast belongs to me [Chorus] We rise we rise we rise Above the cloudy skies Watch me transcend, watch me ascendGeneric tags. Cliché lines. No vocal lock. No BPM. Suno guesses and you re-roll eight times.
Style: melodic rap, 808-heavy, atmospheric Toronto sound, male vocal, mid-tenor, light Auto-Tune, introspective delivery, 72 BPM, minor key, half-time drums, sparse keys, R&B-leaning bridge [Intro] [Verse 1] Two AM and the city sounds dull through the glass I keep one eye on the door and the other on the past You said you’d call back when the rain finally cleared It’s been forty-eight hours and the silence is real [Pre-Chorus] [Chorus] [Verse 2] [Bridge] [Outro]Vocal identity locked. Genre-specific tags Suno actually triggers on. Cliché firewall caught “rise from ashes” before it reached the page. Structural tags placed for BPM. One pass.
Workflow tax, removed
The pipeline handles all six. One pass. Every track.
What producers are doing with it
Live numbers from inside the app. Nothing rounded, nothing inflated.
From producers shipping with HookGenius
“A terrific program to help set my lyrics up for Suno… my songs come out as I imagined them.”
J.D. · HookGenius user
Verified via in-app feedback
“I’m using it every day. It’s part of my workflow for every song. I had a chorus I’d been trying to finish for twenty years — I used the HookGenius lyrics as is. Beyond thrilled at the results.”
R. · HookGenius user
Verified via in-app feedback
Verbatim excerpts from real HookGenius customer messages. Ellipses mark cuts; no words changed. Initials only — full names redacted by default because we don’t publish customer identities without explicit permission.
Pipeline is platform-agnostic
The output format is Suno-native, but the structure travels. Paste it into any prompt-based AI music tool.
v5 and v5.5 metatags, pipe syntax, spacer lines, pronunciation hints — all tuned in.
Copy-paste compatible. Style + structural blocks translate cleanly.
Universal prompt format. Structural tags become section guidance.
Lyrics and style fields map directly.
Frequently asked, plainly answered
No. HookGenius writes the lyrics, the style prompt, the title, the structural metatags, and the vocal direction. The audio render happens inside Suno, Udio, Riffusion, or whichever AI music platform you use. HookGenius is the step before the audio model — the pre-production engine, not the audio engine.
Yes — for generating audio. Suno, Udio, Riffusion, or Mureka handles the audio render. HookGenius makes that render dramatically better by giving the audio model a properly engineered prompt, a Suno-formatted lyrics block, and a vocal identity that actually holds.
HookGenius runs a multi-stage Claude Sonnet pipeline purpose-built for prompt-based AI music. Each stage handles one decision: vocal identity (gender, texture, register locked together), Suno-specific metatag selection across 74 genre profiles, structural tagging with pipe syntax and spacer lines, a cliché firewall at the tonal layer, and album-level cohesion across multi-track projects. A general-purpose LLM can write lyrics. It does not enforce these constraints as a contract.
Producers report a dramatically higher hit rate per generation — tracks that sound finished on the first paste rather than the fifth. HookGenius has not yet published a precise percentage because no controlled side-by-side measurement has been run. The lift is qualitative but consistent across user feedback.
Suno is first-class — the metatags, pipe syntax, and spacer dot lines are tuned to Suno’s exact conventions (v5 and v5.5 today). The output is copy-paste compatible with Udio, Riffusion, and Mureka. The format travels: when a new prompt-based AI music platform ships, the HookGenius output will work there too.
Yes. Five free generations on sign-up, no credit card. After that, credit packs start at $7. There is no subscription requirement.
Three things per generation: a song title, a fully-formatted lyrics block with structural metatags and vocal cues, and a Suno-ready style prompt. Paste the lyrics into Suno’s lyrics field, the style into Suno’s style field, and press Create.
Yes — for the first five generations. The free tier is the full product, same Claude Sonnet pipeline, same Suno-ready output as paid users, capped at five tracks.