Introduction

​

What is Voice Studio?

PeachDesk Voice Studio is Bibha’s hosted AI voice studio, built right into the PeachDesk dashboard. It brings professional voice cloning, text-to-speech, and audio production into a single web app — no installs, no local models, and no GPU of your own required. Everything runs on PeachDesk’s servers. You select an engine and a voice from a dropdown, type or paste your script, and PeachDesk handles the heavy lifting on our GPU infrastructure. The result streams back to your browser, ready to preview, edit, and download. Voice Studio covers the full creative loop:

• Clone a voice — upload a short reference sample and PeachDesk builds a reusable voice profile you can generate with again and again
• Generate speech — turn any text into natural, expressive audio in dozens of languages
• Shape the result — layer post-processing effects, stitch multi-speaker conversations, and export production-ready audio

It’s the hosted alternative to tools like ElevenLabs — voice cloning, TTS, and audio editing in one place, available to your whole team inside PeachDesk.

​

Getting started

​

What’s in Voice Studio

The Studio is organized into tabs across the top of the app:

• Generation — the main workspace: choose an engine and voice, enter your script, and produce audio. Long scripts are auto-chunked and crossfaded so there’s no practical length limit.
• Voices — your voice profiles, both Built-in voices and voices you’ve Clone[d] from audio. Zero-shot cloning works from a single short reference sample.
• Effects — post-processing such as pitch shift, reverb, delay, chorus, compression, and filters. Effects are applied as versioned takes, so the original render is always preserved.
• Stories — a multi-track timeline for assembling conversations, podcasts, and dialogue across multiple voices.
• Talk — speak naturally and have your speech transcribed, then play it back in any of your voices.
• Captures — a paired audio + transcript archive. Retranscribe, refine, or promote a capture to a reusable voice sample.
• Settings — manage your workspace preferences and defaults.

​

Voice engines

Voice Studio offers several text-to-speech engines, each with different strengths. Pick the one that fits your project from the engine dropdown in the Generation tab — all of them run on PeachDesk’s GPU infrastructure, so there’s nothing to download or configure. All of them belong to VAANI, Bibha’s own text-to-speech family:

​

Expressive speech

Some engines let you go beyond flat narration:

• Paralinguistic tags — VAANI understands inline tags like [laugh] and [sigh] to add lifelike, non-verbal sounds. The other VAANI engines have no tag tokens and would read them out literally.
• Fine-tune controls — every engine exposes an expressiveness knob and a sampling temperature; the full-size variants add a pacing (CFG) control. Pick them from the Fine-tune panel next to the generate box.
• Post-processing — pitch, speed, volume, reverb, EQ and more are applied after synthesis on the Effects tab, so they work with every engine.

​

Use cases

​

How it works

You never touch a model file. When you generate, Voice Studio sends your script and voice selection to PeachDesk’s backend, where the chosen engine runs on our GPU servers. The rendered audio is returned to your browser and saved to your workspace. Cloning works the same way: you upload a reference sample, and PeachDesk builds and hosts the voice profile for you.