It started at the hairdresser. The music in the background was an AI-generated lo-fi stream from YouTube, the kind that runs for hours in shops, salons and cafes and quietly collects millions of views. I wanted my own: a playlist of the genres I actually like, and maybe a channel to put it on. Generating songs one by one in the Suno app would never scale to that. So I built a pipeline that does it for me.
What the workflow does
Suno makes remarkably good music from a text prompt, but the app itself has no API. The workaround is Kie.ai, which exposes the Suno model through a normal REST API that n8n can call. From there the system is a loop you can run as often as you like:
- A form kicks it off. You type the genre, the mood, how many tracks you want, and whether they should have lyrics or be instrumental.
- A lyrics agent writes the words. An LLM agent with a structured output parser produces a title, a style description and verses that fit the genre, so every track gets a proper song structure rather than a random prompt.
- A music agent turns that into a Suno request. The HTTP node sends the style and lyrics to the API and gets back a task ID.
- Wait, check, repeat. Generation takes a minute or two, so the workflow waits, polls the status endpoint and loops until the song is ready. No babysitting.
- Extend when the track is too short. A second branch calls the extend endpoint so a two-minute song becomes a four-minute one that works in a playlist.
- Download and file. The finished MP3s land in a Google Drive folder, named and ready to upload, with two variations per generation so you can pick the better take.
Get the complete Suno music workflow from the video
The exact 44-node n8n scenario: lyrics agent, Suno generation through Kie.ai, status polling, song extension and Google Drive upload. Import it, add your keys, run. Leave your email and the file unlocks right away.
- Form trigger: genre, mood, track count, lyrics or instrumental
- Lyrics agent with structured output: title, style description and verses that fit the genre
- Music generator agent + Suno API calls: create, check status, download, two variations per run
- Wait and poll loop: handles the asynchronous generation without manual checks
- Extend branch: turns short tracks into playlist-length songs
- Google Drive upload: finished MP3s filed and named, ready for YouTube
You're in. Here are your files
Unzip, then in n8n choose Import from file and pick the JSON. You will need a Kie.ai API key (for Suno), an OpenRouter or OpenAI key and Google Drive credentials.
Download the files (10 KB)Want this set up inside your company? Book a free consultation.
What came out of it
In the video the same scenario produces a rap track, a blues afrohouse groove, a drum and bass piece, a rap/R&B song and a set of instrumental background tracks for a playlist. Each one took a form submission and a few minutes of waiting. Running it for 40 or 50 tracks is the same effort as running it once.
What it costs
Through Kie.ai a song costs a few cents, so a batch of 50 tracks comes to a couple of dollars. If one playlist on YouTube brings in a fraction of the views the hairdresser's stream does, that pays for itself many times over. Commercial use rights follow the paid Suno plan, so check that before you publish.
Why this matters beyond music
Strip away the music and this is the pattern behind most AI automations we build: an agent prepares a structured brief, an API does the heavy lifting, the workflow waits for an asynchronous job, checks the result and files it where the team already works. Swap Suno for a transcription service, an image generator or a document pipeline and the skeleton stays the same.