BlockTunes Studio: Six Professional Tools, Running on Our Hardware

So a major dream is coming true over at Blocktunes. As a classically and digitally trained and certified audio engineer, I (@thelogicaldude - @groove-logic), have dreamed for years of building automated tools that are actually good. We've had things like LANDR for a long time, but they are not really that great. Well, now with the technology we have at our fingertips has changed the game for me in a big way, because now, my dreams have become reality!

Most platforms that offer AI tools are doing one of two things. They are either reselling access to OpenAI's API at a markup, or building a thin wrapper around someone else's corporate model and calling it a product.

BlockTunes is doing neither. Every AI tool on the platform runs on hardware we own, in a facility we operate, using open source models we have deployed and specifically trained ourselves. No third party API fees. No external dependency that can go down, get rate limited, or change pricing on us. Our tools are our infrastructure. The only monthly bills we pay are power and internet.

Six tools are live right now in the Studio. Here is what each one does.

LyricalGPT

LyricalGPT is our AI chat interface specialized entirely for songwriting. Not a general purpose chatbot. It knows song structure, verse and chorus and bridge form, rhyme schemes, syllable weight, cadence, genre conventions, and the difference between a hook and a pre chorus.

Quick Actions let you generate a verse, chorus, bridge, hook, or full song outline in one click. Genre and Mood chips tune the voice before you start. Responses stream word by word so you can see the direction and redirect before it finishes. Chat history is saved automatically. You can paste in your own lyrics as style references so the model writes in your specific voice, not a generic one.

LyricalGPT is free for every logged in user: 10 messages per day at no credit cost, then 0.5 credits per message beyond that. Well, free users get 2 free credits a day, so that's where the 10 uses come into play. You can honestly use those credits however you want. There is a dedicated post covering everything it can do. The short version: it is 2am, the track is almost done, and you need the right words. It is there to help you out!

TuneEngine

Generate a complete, original track from a text prompt.

Describe the genre, mood, instrumentation, tempo, and style. TuneEngine builds a full track from that description. Not a loop, not a stem, not a fragment. A complete piece of music, ready to listen to and download. From there, you can send it to the Mastering Engine to polish it up for distribution. After mastering, you can go straight to mint or download it and send it to a mainstream distribution service like DistroKid.

The model running under the hood is ACE Step, running on the BlockTunes Mac Mini AI server. No external music generation API. No subscription to a third party service passing through to you. The generation runs on our hardware. That being said, resources are limited, so if it's taking a bit longer than normal, just be patient, it may be busy. As we grow, hopefully we will be able to expand our infrastructure to be able to handle more jobs. We were able to move the mastering engine to another machine because it is less reliant on GPU and the TuneEngine.

How it works:

Type a prompt. Choose whether you want one focused output or two variations to pick from. Hit Generate. The track builds in real time with a progress indicator. When it is done, play it, download it, or hit "Master this Track" to send it directly to the AI Mastering tab with no upload needed. It lands pre loaded and ready to process.

That pipeline goes: Generate, then "Master this Track", then Mastering, then "Mint this Master", then Mint Master. Full creation to published release without leaving the Studio.

TuneEngine is for producers, beatmakers, songwriters needing a demo bed, artists who want to hear a direction before committing to a full session, and anyone who has ever had a concept in their head that they could not execute because they did not have the right collaborator in the room.

Credit cost: 10 credits per generation, whether you generate 1 track or 2.

AI Mastering

Upload your track, choose a genre style, and get a professionally mastered version back in minutes.

Most AI mastering tools are either a loudness maximizer with a nice UI, or a generic reference matching algorithm that treats every track the same. The BlockTunes mastering engine works differently. It listens to your track before it touches it, and the goal is not to make your track as loud as possible. The goal is to make it as clean, balanced, and translatable as possible.

The Analysis First Approach

Before any processing begins, the engine analyzes your specific track. It measures the spectral balance across the bass, mid, and high frequency bands using calibrated targets derived from professionally mastered reference tracks in each genre as well as genre based settings built on results from a test pool of real human audio engineers that we have known and trusted for years. It also measures your track's dynamic range. Based on what it finds, it applies corrective EQ to fix imbalances specific to your material, and scales the compression intensity to match how dynamic or compressed your source already is.

This matters because over compression and aggressive limiting are the most common mistakes in mastering. Every major streaming platform, Spotify, Apple Music, YouTube, applies its own loudness normalization when it plays back your track. If you hand them a brick walled, over limited file, the normalization will turn it down anyway, and what you are left with is a track that lost its dynamics for nothing. The BlockTunes mastering engine targets the right loudness for your genre and stops there. Your track will hit where it needs to hit and breathe where it needs to breathe.

The Full Processing Chain

1. Adaptive spectral analysis
2. Corrective EQ, specific to your track's imbalances
3. Genre character EQ, the tonal signature of the style
4. Air shelf, presence and clarity on top
5. Three band multiband compression with independent compressors for bass, mids, and highs, each with threshold, ratio, attack, and release values tuned per genre
6. Mid side stereo width processing
7. LUFS loudness normalization to the genre's streaming target
8. True peak limiting

Every step is tuned per genre. Lo-Fi and Singer/Songwriter masters stay open and dynamic. Drum and Bass and Metal hit hard where the genre demands it. Hip-Hop and Trap locks the low end. The engine knows the difference. As a professional mastering engineer, this is the one I am most proud of. The rest were really just bonus builds because we could.

26 genre presets:

Pop, Rock, Metal, Hip-Hop/Trap, Electronic, House, Dubstep, Drum and Bass, Lo-Fi, R&B, Jazz, Singer/Songwriter, Afrobeats, Ambient, Classical, Country, Dancehall, Drill, Folk, Gospel, Latin, Orchestral, Phonk, Punk, Reggae, Soul.

After your master is complete, you can preview the original and mastered versions side by side, download the WAV, or send it straight to Mint Master. The mastered file carries over automatically.

Credit cost: 4 credits per master.

Stem Separation

Upload a mix and get back your choice of 4 or six individual stems: vocals, bass, drums, guitar, piano, and other instruments.

This is useful in more situations than you might think.

Remixers who want to work with just the vocals or just the instrumental from a track they produced. Producers who want to revisit a finished mix and pull out individual elements. Live performers who need a backing track with the lead vocal removed. Collaborators who want to add a new layer to an existing arrangement without reworking the whole mix. Anyone sampling their own catalog and wanting clean, isolated source material without digging up old session files.

The model handles real world mixes. It does not need a pristine studio recording to do useful work.

Credit cost: 4 credits per separation.

Vocal Booth

Record your voice or upload any audio file and get back a processed vocal that sounds like it was tracked in a controlled environment. This one came from a feature request from a community member, @elementm. This tool is proof that we listen to our community.

Designed for artists recording outside ideal acoustic conditions, like a home setup with room noise, a phone mic, or a voice memo captured on the spot. The Vocal Booth runs your audio through a multi stage processing chain that adds definition, depth, and polish without touching pitch or timing. This uses some of the same processing tools from the mastering engine, so it was an easy addition to the suite of tools.

The processing chain:

1. Adaptive noise reduction that removes background hiss, HVAC rumble, and room noise without softening the vocal
2. High pass filter at 80Hz to clear low end buildup and handling noise
3. Low pass filter at 14kHz to roll off harsh digital edge above the musical range
4. Voice adaptive EQ where the system detects whether the voice is male, female, or high range and applies warmth and presence boosts tuned to that specific frequency range
5. Gentle 3:1 compression that evens out dynamic swings with attack and release tuned for natural vocal feel
6. Loudness normalization to -14 LUFS

Two input modes:

Record mode: Tap the microphone button and record directly in your browser. Works on phone, tablet, or desktop. No app download, no routing setup.

Upload mode: Drag and drop any audio file, MP3, WAV, M4A, OGG, up to 200MB. Useful for cleaning up existing recordings or processing a phone take you already have.

Credit cost: 2 credits per session.

AI Cover Art

Generate original album and single cover art from a text prompt. Describe the aesthetic, mood, genre, and visual concept. Get something unique back.

No stock photo hunting. No Canva template that half your peers are using. No licensing questions. Describe what you want and the model generates it. "Dark ambient cinematic landscape" gives you something fundamentally different from "bright lo-fi bedroom pop collage." You are prompting from scratch, not choosing from presets. You can also enhance those prompts with our AI prompt enhancement tool.

The generation model runs on the BlockTunes GPU server. No external image API. Now, to be completely honest, this tool is the weakest of the stack. The current GPU in the Blocktunes server machine isn't what we want right now, so this is the first hardware upgrade to be done when we can afford it. So we are running a lower model of Stable Diffusion. Once we are able to update the GPU in this machine, we will be able to get MUCH better results. The foundation has been built, it just needs to expand.

Credit cost: 2 credits per generation.

How to Access the Tools

All six tools are live and accessible from the Studio at app.blocktunes.net/studio.

Access is tier based, with four paths to qualify:

1. Stake $MUSIC: stake enough tokens and your tier activates automatically
2. Delegate Hive Power to @blocktunes: same tier access, your HP stays yours, and you earn passive BT Credits daily
3. Monthly HBD subscription: set up from your wallet, cancellable any time
4. BT Credits: buy a bundle, pay per use, credits never expire

Credits draw from a shared daily pool. Use them on any tool in any combination.

Tool costs: LyricalGPT 0.5 credits per message with 10 free per day. Mastering 4 credits. Stems 4 credits. TuneEngine 10 credits. Cover Art 2 credits. Vocal Booth 2 credits.

The highest tier you qualify for across all paths is what you get.

HP delegators earn BT Credits passively on top of tier access. Every account that delegates to @blocktunes accumulates 0.005 BT Credits per HP delegated per day, paid out nightly. 1,000 HP delegated earns 5 credits per day on top of your daily allowance. Those credits never expire.

New accounts get 10 free BT Credits on signup. No purchase or delegation required.

Buying BT Credits

If you need some extra credits to get the job done, you can purchase a bundle of extra credits that will stay in your wallet until you use them. They are added to your daily tier total if you are a tiered member, meaning your daily credits are used first, extras are used when those run out.

Credit bundles are available at app.blocktunes.net/credits.

Payment options:

We are offering various crypto payment options for BT Credit bundles. All payment offerings use our own hosted RPC nodes for those chains, meaning no middlemen, direct to blockchain payments and confirmations.

MUSIC, HIVE, HBD, Bitcoin, Litecoin, Dogecoin, Ethereum, USDT, USDC, DAI, and Monero. Every chain settles against our own full nodes with no third party payment processor in the middle.

Paying with MUSIC burns those tokens. When you purchase credits with $MUSIC, the platform receives them and immediately transfers them to @null, Hive's permanent burn address. The burn is on chain and verifiable. It is the most direct way to reduce the circulating supply while getting something useful in return.

Credits Are On Chain

BT Credits are not a points system living in a database. Every credit event, every grant, purchase, tool deduction, and peer to peer transfer, is recorded on the Hive blockchain as a verifiable custom_json operation.

You can independently verify every credit transaction ever recorded on the platform at the Credit Ledger Explorer. Filter by any username. Every entry links directly to the on chain transaction.

Credits are also transferrable. Send BT Credits to any other BlockTunes user from the Studio or Credits page. The transfer signs with your Hive posting key and confirms on chain immediately.

About the BlockTunes Cloud

Every AI job runs on hardware we own, in a facility we operate. No AWS, no Google Cloud, no Azure. The mastering engine is ours. The image model is ours. The music generator is ours. The stem separator is ours. We built the stack, we run the stack, we control the quality.

If a job gets stuck or something is not behaving as expected, Harmony, the BlockTunes AI support assistant, is in the bottom right corner of every page. She has your account state and can tell you exactly what is happening. If it needs a staff member, hit Submit Ticket in the chat to open a private thread in the BlockTunes Discord.

Sign Up Free
Open the Studio
Buy Studio Credits

Posted using BlockTunes