Whispering: A Truly Transparent Open-Source Speech-to-Text Solution for Everyday Use

Have you ever found yourself wishing you could effortlessly convert your spoken words into written text? Whether you’re taking meeting notes, brainstorming ideas, or simply trying to capture thoughts on the fly, speech-to-text technology has become an essential tool in our digital lives. Yet, most solutions available today come with significant drawbacks: high costs, questionable privacy practices, and frustrating limitations.

What if there was a tool that let you speak freely while respecting your privacy and your wallet? That’s exactly what Whispering delivers—a genuinely open-source, transparent, and efficient speech-to-text application that puts you in control.

Why Speech-to-Text Tools Often Disappoint

Let’s be honest: most speech-to-text applications fall short in critical ways that matter to everyday users.

Many popular services charge premium prices—typically $15-30 per month—while the actual cost of the underlying technology is just a fraction of that. These middlemen position themselves between you and the service providers, adding unnecessary costs while claiming to offer “local” or “on-device” processing that’s anything but transparent.

The bigger concern? Privacy. When you speak into your device, where does that audio actually go? With closed-source applications, you’re forced to trust a black box with your voice data—data that could contain sensitive information about your work, health, or personal life.

As someone who’s relied on various transcription tools over the years, I’ve experienced this frustration firsthand. That’s why I was particularly intrigued when I discovered Whispering—a tool built from the ground up with transparency as its core principle.

What Makes Whispering Different?

Whispering isn’t just another speech-to-text application. It represents a fundamental shift in how these tools should work:

1. You press a keyboard shortcut
2. You speak your thoughts
3. Your words transcribe instantly
4. The text automatically copies to your clipboard

This simple workflow—press shortcut → speak → get text—delivers exactly what you need without unnecessary complications. But the real magic lies beneath the surface.

Unlike most applications that act as middlemen, Whispering connects you directly to transcription service providers using your own API keys. Your audio travels straight from your device to the provider of your choice—whether that’s Groq, OpenAI, ElevenLabs, or a local service—with no intermediary servers. This means:

• No data collection by Whispering’s developers
• No hidden costs from middlemen taking their cut
• Complete transparency about where your data goes

The creator of Whispering put it perfectly: “I really like hands-free voice dictation. For years, I relied on transcription tools that were almost good, but they were all closed-source. Even those claiming to be ‘local’ or ‘on-device’ were still black boxes that left me wondering where my audio really went. So I built Whispering. It’s open-source, local-first, and most importantly, transparent with your data.”

Understanding the True Cost of Speech-to-Text

One of the most compelling aspects of Whispering is how dramatically it reduces your costs compared to traditional services. Let’s break down exactly what you’d pay:

Service Provider	Cost per Hour	Light Use (20 min/day)	Moderate Use (1 hr/day)	Heavy Use (3 hrs/day)	Traditional Tools
distil-whisper-large-v3-en (Groq)	$0.02	$0.20/month	$0.60/month	$1.80/month	$15-30/month
whisper-large-v3-turbo (Groq)	$0.04	$0.40/month	$1.20/month	$3.60/month	$15-30/month
gpt-4o-mini-transcribe (OpenAI)	$0.18	$1.80/month	$5.40/month	$16.20/month	$15-30/month
Local Transcription	$0.00	$0.00/month	$0.00/month	$0.00/month	$15-30/month

The difference is staggering. With Whispering, you pay only for the actual service usage—typically just pennies per hour—rather than subsidizing a middleman’s profits. The developer reports using Whispering for several hours daily at a total cost of about $3 per month.

This cost efficiency isn’t just about saving money; it’s about eliminating unnecessary layers between you and the service you’re actually using. When you use Whispering with Groq (the developer’s preferred option), you’re paying Groq directly for their service at their published rates, not an inflated price set by an intermediary.

Getting Started with Whispering: A Simple Two-Minute Setup

One of Whispering’s strengths is how quickly you can get up and running. The entire setup process takes about two minutes and consists of three straightforward steps.

Step 1: Download Whispering for Your Operating System

Whispering supports all major desktop platforms with native applications optimized for each system.

For macOS Users

Download Options:

Architecture	Download	Requirements
Apple Silicon	Whispering_7.3.0_aarch64.dmg	M1/M2/M3/M4 Macs
Intel	Whispering_7.3.0_x64.dmg	Intel-based Macs

Not sure which Mac you have?

1. Click the Apple menu → About This Mac
2. Look for “Chip” or “Processor”:

   • Apple M1/M2/M3/M4 → Use Apple Silicon version
   • Intel Core → Use Intel version

Installation Steps:

1. Download the .dmg file for your architecture
2. Open the downloaded file
3. Drag Whispering to your Applications folder
4. Open Whispering from Applications

Troubleshooting Tips:

• “Unverified developer” warning: Right-click the app → Open → Open
• “App is damaged” error (Apple Silicon): Run xattr -cr /Applications/Whispering.app in Terminal

For Windows Users

Download Options:

Installer Type	Download	Description
MSI Installer	Whispering_7.3.0_x64_en-US.msi	Recommended Standard Windows installer
EXE Installer	Whispering_7.3.0_x64-setup.exe	Alternative installer option

Installation Steps:

1. Download the .msi installer (recommended)
2. Double-click to run the installer
3. If Windows Defender appears: Click “More Info” → “Run Anyway”
4. Follow the installation wizard
5. Whispering will appear in your Start Menu when complete

For Linux Users

Download Options:

Package Format	Download	Compatible With
AppImage	Whispering_7.3.0_amd64.AppImage	All Linux distributions
DEB Package	Whispering_7.3.0_amd64.deb	Debian, Ubuntu, Pop!_OS
RPM Package	Whispering-7.3.0-1.x86_64.rpm	Fedora, RHEL, openSUSE

Quick Install Commands:

AppImage (Universal):

wget https://github.com/epicenter-so/epicenter/releases/download/v7.3.0/Whispering_7.3.0_amd64.AppImage
chmod +x Whispering_7.3.0_amd64.AppImage
./Whispering_7.3.0_amd64.AppImage

Debian/Ubuntu:

wget https://github.com/epicenter-so/epicenter/releases/download/v7.3.0/Whispering_7.3.0_amd64.deb
sudo dpkg -i Whispering_7.3.0_amd64.deb

Fedora/RHEL:

wget https://github.com/epicenter-so/epicenter/releases/download/v7.3.0/Whispering-7.3.0-1.x86_64.rpm
sudo rpm -i Whispering-7.3.0-1.x86_64.rpm

Note: If download links aren’t working, visit GitHub Releases for the latest version.

Step 2: Get Your API Key

To connect Whispering to a transcription service, you’ll need an API key. The developer personally recommends Groq for most use cases:

“Why Groq? The fastest models, super accurate, generous free tier, and unbeatable price (as cheap as $0.02/hour using distil-whisper-large-v3-en)”

Here’s how to get started with Groq:

1. Visit console.groq.com/keys
2. Sign up for an account
3. Create an API key
4. Copy your new key

The best part? You don’t need to provide credit card information to access Groq’s free tier. You can start transcribing immediately with no financial commitment.

Step 3: Connect and Test

Now that you have Whispering installed and your API key ready, it’s time to connect everything:

1. Open Whispering
2. Click Settings (⚙️) → Transcription
3. Select Groq → Paste your API key where it says “Groq API Key”
4. Click the recording button (or press Cmd+Shift+; on macOS / Ctrl+Shift+; on Windows/Linux) and say “Testing Whispering”
5. Your transcribed text should now be in your clipboard—paste it anywhere to verify!

If you encounter any issues during setup, don’t worry. The most common problems and their solutions include:

• No transcription? → Double-check your API key in Settings
• Shortcut not working? → Bring Whispering to the foreground
• Wrong provider selected? → Check Settings → Transcription

For platform-specific issues, the documentation provides detailed troubleshooting guides, including solutions for accidentally rejecting microphone permissions or dealing with macOS App Nap (which can suspend background apps to save battery).

Unlocking Advanced Features: Taking Whispering to the Next Level

Once you’ve mastered the basics, Whispering offers several powerful features that can transform how you work with speech-to-text technology.

Multiple Transcription Service Options

Whispering gives you the flexibility to choose from several transcription providers based on your specific needs:

• Groq (Recommended): Fastest models ($0.02/hr), super accurate, generous free tier
• OpenAI: Industry standard models like whisper-1 ($0.36/hr)and‘gpt-4o-mini-transcribe‘($0.18/hr)
• ElevenLabs: High-quality voice AI with models like scribe_v1
• Local Providers (Speaches): Complete privacy, offline use, free forever

This flexibility means you can optimize for speed, accuracy, privacy, or cost depending on your current task. Need maximum privacy for sensitive content? Switch to local transcription. Need the fastest turnaround for a time-sensitive project? Groq’s models deliver remarkable speed.

AI-Powered Text Transformations

One of Whispering’s most powerful features is its ability to automatically transform your transcribed text through customizable AI workflows. Here’s how to set up a basic text formatting transformation:

1. Go to Transformations (📚) in the top bar
2. Click “Create Transformation” → Name it “Format Text”
3. Add a Prompt Transform step:

   • Model: Claude Sonnet 3.5 (or your preferred AI)
   • System prompt: Detailed formatting guidelines (see below)
   • User prompt: Here is the text to format: {{input}}

The system prompt can include comprehensive instructions like:

“You are an intelligent text formatter specializing in cleaning up transcribed speech. Your task is to transform raw transcribed text into well-formatted, readable content while maintaining the speaker’s original intent and voice.

Core Principles:

• Preserve authenticity: Keep the original wording and phrasing as much as possible
• Add clarity: Make intelligent corrections only where needed for comprehension
• Enhance readability: Apply proper formatting, punctuation, and structure

[Additional detailed formatting guidelines would follow]”

These transformations can:

• Automatically fix grammar and punctuation
• Translate text to other languages
• Convert casual speech to professional writing
• Create summaries or bullet points
• Remove filler words (“um”, “uh”)
• Chain multiple processing steps together

For example, you could create a workflow that takes your speech → transcribes it → fixes grammar → translates to Spanish → copies to clipboard, all with a single keyboard shortcut.

Voice Activity Detection (VAD)

If you prefer truly hands-free operation, Whispering’s Voice Activity Detection feature is perfect for you. Instead of holding down a button while you speak, VAD automatically starts recording when you begin speaking and stops when you pause.

Two ways to enable VAD:

1. On the homepage, click the “Voice Activated” tab (next to “Manual”)
2. Go to Settings → Recording → Select “Voice Activated” in the Recording Mode dropdown

How it works:

• Press shortcut once → VAD starts listening
• Speak → Recording begins automatically
• Stop speaking → Recording stops after a brief pause
• Your transcription appears instantly

This feature is ideal for dictation scenarios where you need to keep your hands free—whether you’re cooking, moving around your office, or simply prefer a more natural speaking experience.

Custom Keyboard Shortcuts

Whispering lets you customize the recording shortcut to whatever feels most natural for your workflow:

1. Go to Settings → Recording
2. Click on the shortcut field
3. Press your desired key combination
4. Popular choices include F1, Cmd+Space+R, or Ctrl+Shift+V

This level of customization ensures that Whispering integrates seamlessly into your existing workflow rather than forcing you to adapt to its requirements.

Privacy and Data Handling: Understanding What Happens to Your Information

For many users, privacy is the most critical consideration when choosing a speech-to-text application. Whispering takes a transparent approach to data handling that puts you in control.

Local Data Storage

Whispering stores all recordings and transcriptions locally on your device using IndexedDB, a browser-based database technology. This means:

• Your voice recordings never leave your device unless you choose to transcribe them
• Transcribed text remains on your device until you paste it elsewhere
• No cloud storage means no risk of data breaches affecting your content

Direct Data Flow to Providers

When you choose to transcribe audio, Whispering establishes a direct connection between your device and your chosen service provider:

• Your audio travels straight from your device to the provider (Groq, OpenAI, etc.)
• No intermediate servers handle or store your audio
• You use your own API key, so the provider knows the request comes from you

The developer emphasizes: “Your recordings stay on your device in IndexedDB. When you transcribe, audio goes directly to your chosen provider using your API key. No middleman servers. For maximum privacy, use local transcription.”

Analytics and Telemetry

Whispering uses Aptabase, an open-source, privacy-first analytics service, for anonymized event logging. Importantly:

• No personal data is attached to these events
• You can view exactly what events are logged in the analytics.ts file
• You can turn off analytics in settings at any time

This transparent approach to analytics ensures you’re never in the dark about what data might be collected—and gives you complete control over whether to participate.

Frequently Asked Questions About Whispering

How is Whispering different from other transcription apps?

Most apps function as middlemen charging $30/month for API calls that cost pennies. With Whispering, you bring your own API key and pay providers directly. Your audio goes straight from your device to the API with no servers in between, no data collection, and no subscriptions. The code is open source so you can verify exactly what it does.

What technologies is Whispering built with?

Whispering uses Svelte 5 and Tauri, resulting in a tiny application (~22MB) that starts instantly and uses minimal system resources. The codebase is clean and well-documented, making it accessible for developers who want to learn or contribute.

Can I use Whispering offline?

Yes! Use the Speaches provider for local transcription. This option requires no internet connection, no API keys, and provides complete privacy since everything happens on your device.

How much does Whispering actually cost to use?

With Groq (the developer’s preferred option): $0.02-$0.06/hour. With OpenAI: $0.18-$0.36/hour. Local transcription costs nothing. The developer reports using it several hours daily for a total cost of about $3/month.

Is Whispering really private?

Your recordings remain on your device in IndexedDB. When you transcribe, audio goes directly to your chosen provider using your API key—no middleman servers. For maximum privacy, use local transcription.

Can I automatically format the output text?

Yes! Set up AI transformations to fix grammar, translate languages, or reformat text. These transformations work with any LLM provider you choose to connect.

What platforms does Whispering support?

Desktop: Mac (Intel & Apple Silicon), Windows, and Linux. Web: Any modern browser at whispering.epicenter.so.

What if I find a bug?

Open an issue on GitHub. The developer actively maintains Whispering and responds quickly to user reports.

Why Open Source Matters for Fundamental Tools

Whispering represents more than just another application—it embodies a philosophy about the tools we rely on daily. As the developer eloquently states: “I believe that fundamental tools shouldn’t require trusting a black box. Companies pivot, get acquired, or shut down. But open source is forever.”

This perspective is particularly relevant for speech-to-text technology, which handles some of our most personal data—our voices. When you use a closed-source application, you’re forced to trust that:

• Your audio isn’t being stored or analyzed
• The company won’t change its privacy policy
• The service won’t suddenly become paid or disappear

With open-source software like Whispering, you can verify exactly what the application does. You’re not dependent on a company’s promises—you can see the code for yourself or have someone you trust review it.

The developer’s personal experience resonates with many users: “Productivity apps should be open-source and transparent with your data, but they also need to match the UX of paid, closed-software alternatives. I hope Whispering is near that point. I use it for several hours a day, from coding to thinking out loud while carrying pizza boxes back from the office.”

Getting the Most Out of Whispering: Practical Usage Tips

To help you integrate Whispering seamlessly into your daily workflow, here are some practical tips from experienced users:

For Developers

• Use Whispering to dictate code comments or documentation
• Set up transformations to convert spoken descriptions into code snippets
• Pair with local transcription for maximum privacy when working with sensitive code

For Writers and Content Creators

• Use Voice Activity Detection for natural, uninterrupted dictation
• Create custom transformations to match your specific writing style
• Combine with markdown formatting for direct publishing-ready content

For Meeting Professionals

• Record and transcribe key discussion points during virtual meetings
• Use the clipboard history feature to capture multiple ideas
• Set up transformations to create meeting summaries automatically

For Students and Researchers

• Transcribe lecture notes in real-time
• Use local transcription for privacy when working with sensitive research
• Format transcriptions into study guides with custom transformations

The Technical Foundation: What Makes Whispering Work So Well

Whispering’s impressive performance stems from its thoughtful technical architecture:

• Svelte 5: Provides the UI reactivity with an efficient runes system
• Tauri: Enables native desktop performance while keeping the app small
• IndexedDB & Dexie.js: Handle local data storage reliably
• WellCrafted: Offers lightweight, type-safe error handling
• Rust: Powers native desktop features for optimal performance

This combination results in an application that’s not only feature-rich but also remarkably efficient. At just ~22MB, Whispering starts instantly and uses minimal system resources—unlike many bloated alternatives that consume hundreds of megabytes of memory.

The architecture follows a clean three-layer pattern with 97% code sharing between desktop and web versions:

1. Service Layer: Platform-agnostic business logic
2. Query Layer: Reactive data management with caching
3. UI Layer: Clean components with minimal logic

This thoughtful design ensures Whispering remains maintainable, extensible, and performant—qualities that benefit end users through reliability and continuous improvement.

Building Whispering Yourself: For the Security-Conscious

If you’re particularly concerned about security or simply want more control, you can build Whispering from source:

git clone https://github.com/epicenter-so/epicenter.git
cd epicenter
bun i
cd apps/whispering
bun tauri build

The resulting executable will be in apps/whispering/target/release. This process ensures you’re running exactly the code you expect—no hidden surprises. Such is the beauty of open-source software!

Conclusion: A Tool That Respects Your Time, Privacy, and Budget

Whispering represents what speech-to-text technology should be: efficient, transparent, and respectful of your resources. By cutting out unnecessary middlemen and embracing open-source principles, it delivers exceptional value without compromising on privacy or performance.

Whether you’re a developer needing quick code comments, a writer capturing inspiration, or a professional documenting meetings, Whispering offers a solution that works with you—not against you. Its thoughtful design, flexible features, and transparent data handling make it a tool you can trust with your voice.

The developer’s commitment to building something “better than any closed-source alternative” shines through in every aspect of Whispering. As they put it: “The code is open-source because I believe that fundamental tools shouldn’t require trusting a black box. Companies pivot, get acquired, or shut down. But open source is forever.”

In a world where our voices contain increasingly sensitive information, having a tool that respects your privacy while delivering exceptional performance isn’t just nice to have—it’s essential. Whispering proves that open-source software can not only match but exceed the capabilities of proprietary alternatives, all while keeping you in control of your data.

If you’ve been frustrated with existing speech-to-text solutions, give Whispering a try. You might find, as the developer did, that it becomes an indispensable part of your daily workflow—one that you can use with complete confidence in how your data is handled.