AI Music Generator
From an idea to
a song with vocals
Say what the song is about, or give your own lyrics; AI writes, composes and sings it. With Suno, Lyria and MiniMax Music, no instruments or studio needed.
Persian songs made with AI — listen
Both songs were made from this one Persian description, unedited.
ترانه‌ای احساسی با آواز مردانه‌ای گرم، پرطنین و کمی خش‌دار؛ اجرایی صمیمی و پر از بغض که حس دلتنگی و عشق ازدست‌رفته را منتقل کند. ملودی آرام و تأثیرگذار با پیانوی نرم، سازهای زهی لطیف و ضرب‌آهنگی ملایم، فضایی سینمایی و غروب‌زده بسازد؛ اوج آواز به‌تدریج بالا بگیرد و با شور و اندوهی عمیق در ذهن بماند.
از قلب من راه رفتن نگرفت
The model wrote these lyrics from the description and sang them.
The lyrics (Persian)
[Verse 1] هنوز فنجان تو کنار پنجره‌ست لبه‌اش ردِ لبخند تو را نگه داشت صدای ساعت از اتاق می‌گذرد و خانه بی‌تو تا غروب، نفس نداشت [Pre-Chorus] تو رفتی و باران به شیشه رسید من ماندم و حرفی که هرگز نگفتم [Chorus] کاش می‌شد از این فاصله رد شد یک بار دیگر دستت را گرفت این عشق اگرچه سهم ما نشد از قلب من راهِ رفتن نگرفت [Verse 2] آن شال آبی روی صندلی مانده مثل خودت آرام و دور از دسترس من هر غروب اسم تو را آهسته می‌گویم و می‌ریزد دلم از نفس [Pre-Chorus] تو رفتی و پاییز به خانه نشست من ماندم و عکسی که رو به دیوار است [Chorus] کاش می‌شد از این فاصله رد شد یک بار دیگر دستت را گرفت این عشق اگرچه سهم ما نشد از قلب من راهِ رفتن نگرفت [Bridge] اگر صدایم از سکوتت گذشت بدان هنوز جای تو در من هست نه با گله، نه با خواهشِ بازگشت فقط دلم تنگ است، همین و بس
Lyria sample
The same description with Lyria: Persian vocals with a cinematic arrangement.
AI song generator — from an idea to a finished song with lyrics and vocals
DeepFA’s AI song generator makes a complete song: lyrics, melody, arrangement and a singer. In simple mode, write what the song is about and how it should feel; the model writes the lyrics — in Persian too — composes the music and sings it. If you already have lyrics, use custom mode: give the title, the style of music and your words, and the model sings exactly what you wrote. Everything runs in your browser, with no instruments, studio or software.
The tool works with Suno, Lyria and MiniMax Music, and each model has a fixed price shown on its card before you start. Suno makes two different takes per request and sings Persian lyrics with natural pronunciation. Lyria builds structured songs with cinematic arrangements and can take its mood from a picture. MiniMax Music takes the style apart from the lyrics and writes the lyrics itself when you leave them out.
Choose a male or female singer, or switch on Instrumental for music without vocals for videos, adverts, podcasts and games. Every song is kept in your library with its lyrics and cover, ready to star, rename or download. For deeper work — extending a song, separating vocals and instruments, multi-track mixing — use AI Music Pro.
Everything for making songs with AI
From one sentence or your own lyrics to a finished song with a singer
Complete songs with vocals
Lyrics, composition, arrangement and a singer, all from one description, in Persian and other languages.
Simple and custom modes
In simple mode write just the idea; in custom mode give your own title, style and lyrics.
Lyrics written by AI
Get two or three lyric drafts from a short topic and pick one for the song.
Singer, instrumental and length
Choose a male or female singer, music without vocals, and the length you need.
Mood from a picture
Give Lyria a picture and the music follows its mood.
Two takes per request
Suno makes two different takes each time so you can pick the best.
A searchable library with bulk download
Each song is kept with its lyrics and cover; search titles and lyrics, tick several songs and download them as one ZIP.
AI song models
Each model has its strength; make one idea with several and compare
Suno — natural Persian vocals
Two complete songs of up to six minutes per request, with its own lyrics or yours, a male or female singer, and styles to leave out.
Lyria — cinematic arrangements and pictures
Structured songs of a few minutes with detailed instrumentation; it can take its mood from a picture.
MiniMax Music — style and lyrics apart
Takes the style and mood apart from the lyrics; leave the lyrics out and it writes them from your description.
Why choose DeepFA AI Music?
Simple to start, with full control over lyrics, singer, style and length.
Persian songs that sound Persian
Persian lyrics and vocals with natural pronunciation; listen to the samples on this page.
Your lyrics, word for word
In custom mode the model sings the words you wrote, in the structure you marked.
3 models on one page
Make one idea with several models and compare the results side by side.
Clear price, credit back on failure
Each model’s price is known before you start, and the credit comes back if no song is delivered.
From an idea to a song in four steps
No music knowledge, instruments or studio needed.
Choose the model
Each model’s price is written on its card.
Simple or custom
In simple mode write the idea; in custom mode give your title, style and lyrics.
Set the singer and length
Pick a male or female singer or instrumental, and set the length if you need to.
Generate and listen
After one to three minutes the song appears in your list with its lyrics and cover.
Who makes songs with AI?
From video creators and brands to birthday gifts and podcasts
Video and content creators
Soundtracks and custom songs for videos, reels and adverts.
Brands and businesses
Jingles and advertising songs with your brand name and campaign message.
Gifts and occasions
Birthday, anniversary or wedding songs with real names and memories.
Podcasters and game makers
Podcast intros and outros, and instrumental pieces for game levels.
DeepFA AI Music interface
Click any image to enlarge.
How does making a song with AI work?
From writing the description and lyrics to choosing the model and singer — what makes a good AI song.
Song models have learned from millions of pieces of music how melody, harmony, rhythm and vocals fit together. When you describe what a song is about and how it should feel, the model first lays out the lyrics and the structure — intro, verse, chorus, bridge — then composes the melody and arrangement, and finally a singer performs it. Give your own lyrics and the model sets your words to the melody.
Two things shape the result most: a precise style (genre, instruments, tempo, mood) and lyrics with their parts marked. Writing [Verse], [Chorus] and [Bridge] above each part tells the model where to rise and where to settle. For music under a voice, switch on Instrumental.
Writing a good song description
Instead of “a happy song” write “upbeat pop for a birthday, guitar and handclaps, medium tempo, female singer”. Genre, instruments, tempo, mood and singer are the five things that matter most.
Lyrics with marked parts
Write [Verse], [Chorus] or [Bridge] above each part. Repeat the chorus so it sticks, and keep verses short so the singer can breathe.
Choosing the model
Each model has its strength: Suno, Lyria and MiniMax Music. Make one idea with two models and compare; each price is on the model’s card.
Singer or instrumental
For a sung song, choose a male or female voice. For music under narration, a teaser or a game, switch on Instrumental so no vocals are made.
Song length
Set a shorter length for reels and adverts and a longer one for a full song. Leave it unset and the model picks a length that fits the lyrics.
From a good take to a great one
If a result is close, keep the style and generate again — each run makes a new take. To fix just one part of a song, use the section replace tool in AI Music Pro.
For the best AI songs, describe the style precisely, mark the parts of your lyrics, choose the singer and try one idea with two models. The samples at the top of this page show what one good Persian description produces.
More AI Audio & Music Tools
From music creation to voice cloning, voiceover and speech-to-text — all tools together
Which music model should I pick?
How many songs each request makes, how long they run and what each model costs are in the table below.
About 75 credits a song.
Up to 6 min.
Each request makes 2 songs.
| Model | Songs a request | Song length | Cost (a request) |
|---|---|---|---|
|
Suno V6
Cheapest a songLongest
|
2 songs | 10 s–6 min | 150 credits about 30,000–45,000 Toman |
| Lyria 3.5 | 1 song | 30 s–4 min | 250 credits about 50,000–75,000 Toman |
| MiniMax Music 2.5 | 1 song | — | 375 credits about 75,000–112,500 Toman |
This table is built from DeepFA's current models and prices and changes the same day a price changes or a model is added. Toman follows what a credit costs in the plans and packs (about 200–300 Toman); each job's exact cost is shown before it is made, and credit is returned if it fails.
FAQ about making songs with AI
Persian, models, lyrics, timing and credit.
Make your first song now
Say what the song is about or give your lyrics; a few minutes later it is sung and ready.