Voice cloning is on the roadmap and listed as 'coming soon' on the official features page. When shipped, it will apply a consistent vocal style across generations within a project. Until then, vocal output depends on which underlying music model handles the request.

Tunee AI — Chat to Make Music & Cinematic Videos

"AI music tools were getting more sophisticated — and more intimidating. We wanted music creation to feel like a conversation, not a prompt-engineering exam."

The Tunee team

Previously: TemPolor · ChangYa (10M+ MAU)

Where it came from

Built by the people who shipped TemPolor and a 10M-MAU singing app.

Before Tunee, the same team built TemPolor — a royalty-free AI music platform powered by their TemPolor v3.5 model — and ChangYa, a karaoke-style singing app that reached over 10 million monthly active users globally. So when they started Tunee in late 2024, they weren't guessing what creators wanted; they'd been watching it for years.

The pitch is small and stubborn: most AI music tools make you a prompt engineer. Sit, type, regenerate, type again. Tunee flips that into a chat. You describe a song, it drafts. You say "more drums, less reverb," it edits the same draft. It remembers what you liked from earlier in the session.

We'll say the trade-offs out loud. Commercial rights need the paid tier (around $18/mo). The cinematic MV feature leans on third-party models — Midjourney V7, Kling V2.1, Dreamina V3 — so video quality rises and falls with what those models can do that week. Voice cloning is still rolling out. None of that is a deal-breaker, but you should know it before you install.

Reviewed by the Tunee AI editorial desk

Last updated · 30+ chat sessions, 12 MVs, free + paid tier

What's inside · six panels

Most AI music apps are prompt boxes. Tunee is a conversation.

Six things Tunee does that most generators don't even attempt.

01 · The chat

You talk to it. It remembers.

One text box, conversational. Reference a song, hum a melody (audio in), drop a video clip — Tunee reads tone and tempo, remembers earlier turns in the session, and adjusts. No re-prompting from zero.

"warmer pads, less drums"

got it — softening kit, doubling pad layer

▸ draft v2 · 1:48 · stems ready

02 · Multi-model

Switch engines mid-track.

Mureka, Minimax, TemPolor, Ace-Step, Kling — Tunee picks per use case or you toggle manually. Different feel, same project.

Mureka Minimax TemPolor Ace-Step Kling

03 · Stems

Split into 6 stems.

2/4/6-source separation: vocals, drums, bass, melody, FX. Remix, sample, isolate, or pull into your DAW.

VOCAL

DRUMS

BASS

04 · MV gen

Track → 60s cinematic MV.

Story-driven music videos with anime, contemporary, or stylized aesthetics. Powered by Midjourney V7 + Kling V2.1 + Dreamina V3.

05 · Lip-sync

Photo → singing character.

Drop any portrait, Tunee animates it to lip-sync your vocal. Useful for thumbnails, lyric videos, social cuts. No video-editor skills required.

06 · Tunee Skills · integrations

Plug it into your workflow.

Tunee Skills exposes generation as an agent API — ClawHub, GitHub, custom pipelines. For dev teams building music into their own apps without spinning up a model.

Open in app · explore Skills →

How it works · 4 turns

The whole flow fits in one chat thread.

Prompt → draft → refine → MV. No new project, no re-uploading stems, no "start over."

01Describe the song. Free-form sentence. Reference a track, hum a melody (audio in), or drop a video clip — Tunee reads context.

02Get the draft. Tunee picks a model (or you switch manually), composes 1–3 minutes of music, returns stems on demand.

03Refine in the same thread. "More drums," "swap the chorus to minor," "make the bridge longer." Same conversation, same project.

04Generate the MV. Pick a visual style — anime / contemporary / stylized — and Tunee turns the track into a 60-second cinematic video.

Session · Late-Night Drive

Mureka · 78 BPM · D-min

You · 09:42"moody lofi beat for late-night driving, melancholy but warm"

Tunee · 09:4278 BPM, lofi piano, vinyl crackle, soft pads. Drafting…

▸ draft v1 · 1:48 · stems ready · key D-min

You · 09:44"warmer pads, less drums"

Tunee · 09:44softening kit, doubling pad layer. Continuing same project.

▸ draft v2 · 1:48 · 4 stems · MV ready

Honest comparison

Where Tunee wins, and where it doesn't.

Three categories where Tunee loses to incumbents. We're not going to pretend the math works any other way.

Capability	Tunee AI	Suno v4	Udio
Conversational chat refinement	Native, multi-turn	Single prompt, regenerate	Single prompt, regenerate
Music video generation	60s cinematic MV built-in	External tools needed	External tools needed
Switch underlying model	5+ engines, one workflow	Suno-internal only	Udio-internal only
Stem separation	Up to 6 sources	4 stems	4 stems
Voice quality & vocal realism	Solid, model-dependent	Best-in-class vocals	Studio-grade vocals
Genre coverage & consistency	Wide, occasionally uneven	Very consistent	Very consistent
Free tier generosity	Daily refresh, no commercial	Limited daily credits	Limited daily credits
Workflow / API integrations	Tunee Skills · ClawHub, GitHub	Limited public API	No public API

User voices

Three real takes from people who use it daily.

One mixed review left in on purpose. Every AI music tool has one. Read it before you decide.

★★★★★

"I dropped a 15-second clip of the boss fight, asked Tunee for 'something cinematic but angry,' and got a usable track in under three minutes. The MV gen turned it into a trailer I posted that night."

Kai L.

Singapore · indie game dev

★★★★★

"Suno still wins on vocal realism. But for everything around the song — refinement, stems, MV — I'm in Tunee. The fact that I can switch models without leaving the chat is the part I didn't know I needed."

Mara D.

São Paulo · content creator

★★★★☆ honest take

"The MV is hit-or-miss — sometimes amazing, sometimes the lip-sync drifts. Worth saying out loud: it's tied to third-party video models, so quality varies. Music side is solid though."

Jonah R.

Berlin · ad agency

Frequently asked

Questions people ask before downloading.

Is Tunee AI free?

There's a free tier — daily refresh credits, a generous handful of chat-driven generations per day, mood presets, and access to most models. Commercial rights and higher-volume usage require a paid tier starting around $18/month, which gets you ~250 songs/month plus premium stem separation. Pricing can change; check the in-app store rather than relying on a screenshot from a blog post.

How is the chat-based flow different from Suno or Udio?

Suno and Udio are prompt-and-regenerate: you write a prompt, get a result, write another prompt, get another result — every regeneration is a fresh start. Tunee remembers the session: you can say "warmer pads" or "swap the chorus to minor" and it edits the same draft. Same project, multi-turn. Less re-prompting, more conversation.

Which AI models does Tunee actually use?

For music: Mureka, Minimax, TemPolor (their in-house v4 model launched Dec 2025 on the AR + DiT architecture), Ace-Step, plus their own legacy v3.5. For video: Midjourney V7 for images, Kling V2.1 + Dreamina V3 for video synthesis. Tunee picks per task or you can switch manually within the same conversation.

Do I own what I generate?

On the free tier, you can use generations for personal experimentation but not commercially. The paid tiers grant full commercial rights for songs created during your active subscription, plus 1 retroactive commercial right per month for older free-tier tracks. Read the terms in-app for the exact wording; AI music licensing is still legally evolving.

How good are the music videos?

Honest answer: they're hit-or-miss because they lean on third-party video models (Midjourney V7, Kling V2.1, Dreamina V3) which themselves are still maturing. Story-driven and stylized aesthetics tend to land best; tight lip-sync on close-up portraits is the most variable. Quality has been improving sharply month-over-month as the underlying video models update.

Can I clone a voice?

Voice cloning is on the roadmap and listed as "coming soon" on the official features page. When it ships, it'll let you apply a consistent vocal style across multiple generations within a project. Until then, vocal output depends on whichever underlying music model you're routed to.

Who's behind Tunee AI?

The team previously built TemPolor (AI-powered royalty-free music platform, still active) and ChangYa, a karaoke-style singing app that hit 10M+ monthly active users globally. The Tunee v4 model launched in December 2025; they're publicly researching next-gen architectures beyond AR + DiT.

Does it work on iPhone and Android?

Yes — Tunee is available on both iOS and Android via the official store listings, plus a web app at tunee.ai. The mobile and web versions share the same chat history and project canvas, so you can start a session on a laptop and finish it on a phone.

Chat to it. Make a track. Ship the music video.