I Built a Desktop App That Saves 80% on ElevenLabs Costs + Uses Gemini to Polish Scripts

BitProgram — Tue, 11 Aug 2026 09:19:04 +0000

The Problem I Noticed

The more I worked with AI voice generation, the more I saw the same complaint everywhere:

"I keep regenerating the same sentences and burning through credits."

Content creators, YouTubers, audiobook narrators, and course builders were all doing the same thing – writing a script, generating audio, tweaking one word, and generating the entire thing again.

Same text. Same voice. Same emotion. Charged every single time.

Even worse, everyone was wasting credits guessing which emotion sliders to use. Generate. Listen. Adjust. Generate again. Burn more credits.

I realized this wasn't just inefficient – it was expensive and unnecessary.

So I decided to build a solution.

The Solution I Built

I created a desktop app called SoundScript that solves three major pain points:

1. Smart Caching – Generate Once, Play Forever

The app remembers every audio file you've ever generated. If you ever generate the exact same text with the same voice and emotion settings again, it plays instantly from your local storage.

Zero API cost. Zero wait time.

No more paying twice for the same line. No more staring at loading spinners for repeated content.

2. Gemini AI Assistant – No More Guessing Sliders

I integrated Google Gemini directly into the app. Here's how it works:

You paste your raw script or dialogue into the app.
Click the "✨ AI Suggest" button.
Gemini reads your text and automatically recommends:
- The right emotion (e.g., Panic, Joy, Sadness, Neutral)
- The ideal stability and similarity slider settings

No more trial-and-error. No more burning credits testing different emotion combinations. Gemini does the heavy lifting in one click.

3. Batch Processing – Handle Entire Folders

If you have 50 scripts to convert, doing them one by one is torture.

SoundScript lets you drop an entire folder of .txt files and process them all sequentially. The app checks the cache for each one – existing files play instantly, only new ones hit the API.

Who Is This For?

Content creators (YouTubers, TikTokers) generating voiceovers for videos
Audiobook narrators producing long-form audio content
Educators & course creators building voice-driven online courses
Podcasters creating intro/outro voiceovers or full episodes
Game developers generating NPC dialogue
Writers testing how their dialogue sounds aloud
Anyone who uses ElevenLabs regularly and wants to cut their API bill

The Tech Stack (Briefly)

I built this entirely in Python with a clean desktop GUI using CustomTkinter.

The app runs locally on your machine – no cloud storage, no telemetry, no mandatory account. Just a lightweight executable that respects your privacy.

Pricing That Respects You

One-time purchase. Lifetime license.

No recurring fees. No monthly subscriptions. No "credits" to buy. You pay once, use your own ElevenLabs and Gemini API keys, and that's it.

Special Discount for the Community

Since you're reading this on Dev.to, use code Developer26 for 20% off.

Want to Become an Affiliate?

If you're a YouTuber, blogger, or content creator, I offer a 30% one-time commission on every sale you refer.

Email me at bitprogram0@gmail.com if you're interested – I'll set you up with a unique referral link and promotional assets.

Try It Today

Link: SoundScript

Stop paying twice for the same audio. Let Gemini do the guessing. Generate smarter.

DEV Community: BitProgram