REF: MOD_L01 // AI EFFECT CHAIN CONSOLE // VST3 · WINDOWS

Vocant

Describe The Sound. It Builds The Chain.

An AI effect chain console, built into your DAW.

Type what you're after and Vocant reaches for the right tools out of a 32-effect suite and dials them in — in one to three seconds. It's an assistant, not a guide: give it a vibe (“warm tape saturation”) and it makes the call by ear; give it a spec (“reverse reverb at 1/8th”) and it sets exactly that, no interpretation involved. Nothing about your audio ever leaves the room — it runs on your own machine by default.

▸ Watch the 90-second demo

$49 introductory  · $79 at full release

VST3 (Windows) 32 Effect Modules · 283 Parameters Local-First AI 14-Day Free Trial

You Know The Sound.
Finding It Costs You The Take.

// the old way

HABIT_A

Scroll the preset list

Four hundred patches with names like WIDE VOX 3. You audition thirty, two are close, neither is right, and by the time you have edited one into shape the idea you were chasing has gone cold. Presets are somebody else's guess at a song they never heard.

HABIT_B

Build it by hand, again

Compressor, EQ, amp, delay, reverb — five windows, five layouts, and the order matters. You have built this exact chain a hundred times. It is not hard. It is just twenty minutes you spend with a mouse instead of an instrument.

The gap is not knowledge. You can describe the sound in a sentence — that is the fastest, most accurate thing about the whole process. What is missing is something that takes the sentence and turns the knobs, then gets out of the way and lets you move them yourself.

What You Actually Do With It

Type what you want. Tweak what it gives you.

USE_01 // AN ASSISTANT, NOT A GUIDE

Talk to it like a vibe.
Or talk to it like a spec sheet.

Most "AI plugin" gimmicks pick a vague preset and call it done. Vocant listens to how specific you get. Describe a feeling and it makes the tasteful call. Name the exact setting and it stops guessing — it turns the knobs to precisely what you said.

You Say Vocant Does
“reverse reverb” Picks a reverse reverb sound by ear — decay, wet level, and timing all judgment calls.
“reverse reverb at 1/8th” Same effect, but tempo-synced to an eighth note because you said so — the value you named, not its own guess.
“tight vocal compressor” A compressor with fast attack and sensible ratio, tuned by ear for a vocal.
“4:1 ratio, fast attack” A compressor set to exactly that ratio and attack time. Nothing rounded off.
“classic 5150 crunch” Loads a matching amp capture from your own library — see USE_03.

▸ It answers in one to three seconds, and asking for the same thing twice is instant — it remembers what it already worked out.

USE_02 // ONE OF THIRTY-TWO

The AI gets you close.
The knobs are still yours.

Every module in the chain — all 32 of them, from a 6-band EQ to a 16-band vocoder — opens as a full panel: bypass toggle in the header, every control fully named, double-click any value to type it exactly. Drag a module to reorder the chain. The AI gives you a starting point, tuned to what you described; from there it's a plugin like any other.

▸ It's a real plugin underneath, not a black box. Everything the AI sets, you can grab and change by hand — the prompt is a place to start, not a cage.

One of Vocant's 32 effect modules, Tape Stop, showing its Speed, Sync, Div, Stop Ratio and Mix controls with a bypass toggle in the header

FIG_02 // Tape Stop, one of 32 modules — the same panel language across the whole chain

USE_03 // HOW IT DECIDES

It knows the difference between adding and starting over

Every prompt is read against whatever's already active in the chain, so Vocant can tell an addition from a rebuild — and it reaches into your own gear before it reaches for a generic model.

Additive Prompts

“Add chorus” or “also some reverb” keeps everything currently active and layers the new module in — your EQ and compressor stay put.

Replace Prompts

Describe a whole new sound — “clean jazz tone” — and Vocant rebuilds the chain from scratch, dropping whatever doesn't fit.

Your Own Captures

Name a capture you own — “5150 guitar tone” — and Vocant loads that exact .nam file from your own library, not a generic amp sim.

Signal-Flow Order

Multi-effect prompts land in the right order automatically — dynamics, then tone, then time-based — not stacked in whatever order the model thought of them.

▸ Your amp and cabinet captures never leave your drive — Vocant only ever looks in the folders you point it at.

Where It Goes To Work

Three moments a blank effect chain usually costs you time.

Getting Unstuck On An Empty Chain

Staring at a bare insert slot, type what you're going for and get a real starting point instead of scrolling presets that don't quite fit.

Matching Your Own Rig

Name an amp or pedal capture you actually own and Vocant reaches into your own NAM and IR folders for it, not a generic model.

Keeping The Take Moving

Mid-session, type a follow-up — “less reverb, more grit” — instead of hunting through six plugins for the right knob.

// the whole idea

The Model Runs Here,
Or It Does Not Run.

Vocant ships with a local language model and uses it by default. Your prompts, your chain, your captures and your audio stay on the machine in front of you. There is a cloud option for anyone who would rather not run inference locally, and it is off until you switch it on — the whole product works with the network cable pulled out.

That choice costs something and it is worth being plain about it: a local model means a one-time model download and a real dependence on your GPU. On a decent card you get an answer in a second or three. On CPU alone you wait, sometimes twenty-five seconds. We would rather hand you that trade honestly than quietly ship your session audio to a server so the demo feels snappier.

Specifications

// what it is, what it needs, what it costs you in wait

Product Vocant
Version 1.0.0
Format VST3 (64-Bit)
OS Platform Windows 10 / 11 (64-bit)
Download Size 152 MB · ZIP
Effect Modules 32 Modules, 283 Parameters
AI Engine Local LLM (Cloud Optional)
Response Time 1–3s GPU · 8–25s CPU-Only
First-Run Download ~2.4 GB (LLM Model)
Trial Period 14 Days Fully Unlocked
Licensing 3 Machines per Licence
DIST_01 // WINDOWS INSTALLER

Download for Windows

Vocant 1.0.0

152 MB download · ZIP · VST3 · Windows 10/11 64-bit

Download

SHA-256 (extracted .exe): 1614ceab1415cc152947c7006efe806e62806f24fd4e7a156eac6bbf139a53ee

Read the Manual

Unzip, run the installer, and the VST3 lands where your DAW already looks. The first time you open Vocant it fetches the language model — about 2.4 GB, once, and it tells you it is doing it. After that the plug-in runs entirely offline.

LIC_01 // TRIAL

14 Days, Fully Unlocked

All 32 modules, the full parameter set, and the local model. No payment details, no account, no watermark, nothing held back.

When it expires the plug-in stops processing until a licence is activated. Sessions still open and the chains you built are still in them.

LIC_02 // PURCHASE

Buy Once. Keep It.

One-time purchase, perpetual licence. No subscription, no renewal, no per-prompt billing, no tiers. Covers three computers.

$49  Introductory · $79 at full release

Buy Now — $49

Or unlock from inside the plug-in after the trial.

Straight Talk

  • ▸ Windows will warn on the installer. It is unsigned. Click More info → Run anyway. A SHA-256 checksum is published next to the download so you can verify what you got.
  • ▸ There is a 2.4 GB download on first run. That is the language model. It happens once, it is not part of the installer, and you need the connection for it. After that Vocant never needs the network again unless you deliberately turn the cloud option on.
  • ▸ Without a GPU it is slow, not broken. One to three seconds on a reasonable graphics card. Eight to twenty-five on CPU alone. It still gives the same answer — you just wait for it. Try the trial on your own machine before you decide that is acceptable.
  • ▸ It gets close. It is not a mix engineer. A prompt lands you a real starting chain in seconds, not a finished record. Expect to move things afterwards — that is what the 283 parameters are for, and why every one of them is exposed instead of hidden behind the prompt box.
  • ▸ It only knows the captures you own. Ask for a 5150 and it loads a 5150 capture from your NAM library if you have one. If you do not, it reaches for the closest thing it can build from the modules. It does not download gear on your behalf.

Privacy Policy · Terms of Use