finetuning
What's new, improved, and fixed in finetuning. We ship updates every week and share the details here.
This week we kicked off some of the deeper, more technical work to take finetuning to the next level. It's taking a little longer than a week to get live, but it's already in internal testing. We expect to ship early next week, Tuesday or Wednesday, with a further release on Friday.
The Quick Create menu just got 19 brand-new templates — one-tap starting points across beats & hip-hop, electronic, band & vocal, cinematic & world, and global & pop. Each comes tuned with its own tempo, key, and feel. Pick a card on the home page or in Simple mode and generate.
Finetuning now connects to Zapier, wiring your music generation into 8,000+ apps with no code. Three building blocks: a New Completed Track trigger that fires the moment a track finishes (MP3 link, title, duration, settings attached), a Create Music Track action to kick off a track from any Zap, and a Find Track by ID search. Available on Pro and AppSumo Lifetime plans, in private beta. Docs: docs.finetuning.ai/docs/integrations/zapier.
ft — Generate Music From Your TerminalFinetuning now has a CLI. ft lets you generate, list, inspect, and download tracks straight from your shell. Available on Pro and AppSumo Lifetime plans. Docs: docs.finetuning.ai/docs/cli/commands.
ft From Any AI AgentA new skill teaches Claude (and other AI agents via AGENTS.md) how to drive ft for you — ask in plain English and it runs the right commands. Docs: docs.finetuning.ai/docs/cli/skill.
Drop a Finetuning player into any website. Open the share popup on a public track or playlist, switch to the new Embed tab, and pick an inline iframe or a floating widget. Choose one of 12 visual presets (or tweak theme, accent, corners, border, size) and copy the snippet. Available on Plus, Pro, and AppSumo Lifetime plans.
Pick grid or list on the Library page and we'll remember it next time. No more flipping back to grid every visit.
The post-signup welcome email used to say "you have 10 free credits!" which sounded like you could spend them all on day one. Reality: free tier is 10 credits a month, up to 2 tracks a day. The email now says exactly that, so new users aren't surprised when they hit the daily cap.
Public share links — finetuning.ai/t/... for tracks and finetuning.ai/p/... for playlists — no longer require an account just to listen. Hit play and the music starts.
You can now set a password on any public share link, available on every tier including free. Toggle the password on inside the share popup on any of your tracks or playlists, type a password, and only people with the password can play it. Useful for soft-launches, sending a track to a specific group, or keeping a public link technically alive while you decide who's allowed in. Change or remove the password at any time from the same share popup.
A new type of share link, built for creators sending tracks to paying clients, agencies, or collaborators. Open any track's share popup and switch to the Secure tab to create one. No account required for the recipient — they get a clean, private listening page with your branding applied (logo, colors, custom background) for a true white-label feel. Links are expiring (default 7 days, configurable), support optional passwords, show view counts so you can spot leaks, and can be revoked any time without rotating URLs on every link you've sent. Secure Links are available on Pro plans and AppSumo Lifetime Tier 2+.
If you're building on top of the Finetuning public API, you can now skip polling. Add a webhook URL to your POST /v1/generations request and we'll POST the finished track straight to your endpoint the moment it's ready. The callback fires once, identifies itself with User-Agent: finetuning-webhook/1.0, and includes the full generation payload (id, status, audioUrl, prompt, parameters). Add your own secret as a query string on the webhook URL and we'll preserve it on the way out, so you can verify the callback came from your own request. Full guide and examples are on docs.finetuning.ai.
Max track length on paid plans is now 3 minutes 30 seconds (was 3 minutes), both on the web app and via the API. Free tier stays at 2 minutes.
We tightened up the share popup to make it easier to understand at a glance — clearer toggles, less digging around, and a better path to managing your existing links.
You can now add your own prompt on top of a Simple mode Template. Pick a Template for the overall sound, then layer in a short prompt to nudge the vibe, instrumentation, or mood — the best of both worlds between one-click generation and the full prompt experience.
We reworked the parameters in Advanced settings to be more user-friendly — clearer labels, better defaults, and less guesswork about what each knob actually does.
To keep the music generation queue healthy for everyone, the public API now allows 10 generations per minute per user on POST /v1/generations. Reads (status polling, listing, account info) stay on the broader 60/minute limit so you can keep checking on jobs without burning your generation budget.
Studio is here. Take any track you've generated and reshape it without leaving finetuning. Trim lets you pick the exact section you want and download just that piece — bit-perfect quality (no re-encoding) and exports in seconds. Loop turns a short section into a longer track, anywhere from a few seconds up to 60 minutes, with a seamless seam at every repeat. Perfect for background music, study sessions, livestream beds, and content scoring. Open Studio from the home page or jump straight in from any track — the waveform editor lets you scrub, drag the selection, and preview your edit before downloading. Long exports show live progress and an estimated time remaining. Works on desktop and mobile (including iPhone), with desktop recommended for the longest loops. Studio is a premium feature, currently in beta — no watermarks, full commercial rights on everything you export.
We bumped the prompt character limit from 500 to 700 so you have more space to describe the sound you want. The prompt input now also shows a live character counter as you type, so you'll see exactly how much room you have left instead of running into an error after you hit generate.
We tightened how account plan status is refreshed across the app. If you upgrade, redeem an AppSumo code, or change plan level, the create flow now checks your latest account status before deciding whether you can generate. This is especially helpful for AppSumo users who upgraded tiers while already signed in — previously, an old browser session could still behave like a free account until you logged out and back in. Now the app refreshes your plan in the background so your new access is recognized much more reliably.
Not every user wants to write a prompt. We added a Simple mode on the create page for exactly that — pick a curated Template, optionally add lyrics, set the length, and hit generate. That's it. No prompt writing, no mixer, no tag picking. 36 curated Templates to start (with more on the way), each pre-tuned with its own sound, BPM, and musical settings. Quick Create on the home page is now the Simple mode entry point — tap a Template card and you're one step away from a finished track. Advanced mode is still right there when you want the full prompt + mixer experience, and you can switch between the two without losing your place.
The AI lyrics enhancer now has an "Other" language option. Pick it and type your draft in any language you speak — Thai, Japanese, Vietnamese, anything — and the enhancer will work in that same language instead of silently translating everything to English.
On shared track pages, the action buttons (Play, Like, Queue, Copy link) now sit directly under the cover image on mobile so you can hit Play above the fold — no more scrolling past the title, description, stats, and tags to find the controls.
You can now download your tracks as WAV files for higher quality audio. When you hit download, a format picker lets you choose between MP3 or WAV.
The duration slider now goes down to 5 seconds with 1-second precision — perfect for short intros, bridges, jingles, or sound effects. You can also type the exact length directly (supports m:ss format like 0:15 or 1:30). Same goes for BPM — click the number to type an exact value.
Cloning a track now asks you what you want: a new variation (same settings, fresh seed) or an exact clone (identical reproduction, same seed). No more confusion about whether clone gives you a copy or something new.
Tap the three-dot menu on any featured track to see the full prompt and tags used to create it. Copy the prompt or jump straight to the create page with it pre-filled — great for learning what works.
The audio player now has a volume slider. On desktop it's a popup next to the player controls; on mobile it's a slider in the expanded player view. Click/tap the speaker icon to mute/unmute.
Free plan users hitting their generation limit now see a clear upgrade prompt instead of a confusing "too many requests" error.
The prompt field now shows a descriptive example placeholder: "Describe the vibe you want... e.g. chill lofi with acoustic guitar for a rainy day" — helping new users write better prompts from the start.
We added a dedicated section on the features page explaining exactly what Custom Branding includes: themes, logo, background, bio, social links, and white-label options. Plus a new FAQ entry and updated pricing descriptions.