What is VoxKit?

Voice Cloning Text-to-Speech Podcast Generator BYOK

VoxKit is a multi-voice audio studio that lets you create professional podcasts, interviews, audiobooks, and dialogues using AI-generated voices. Record or upload voice samples, clone them into AI voices, write a script with different speakers, and generate a complete audio production — all from your browser, on mobile or PC.

Unlike other voice tools that lock you into one provider, VoxKit supports 4 different AI voice providers — bring your own API key from ElevenLabs, Azure Speech, Google Cloud TTS, or Hugging Face, and use whichever provider offers the best quality for your needs. This BYOK (Bring Your Own Key) model means zero markup on AI costs — you pay the provider directly.

How It Works

1. Add Voices: Create voice profiles for each person (Person A, Person B, Person C, etc.). Record from your microphone or upload an audio sample.
2. Clone Voices: Click "Clone Voice" to send the sample to ElevenLabs — it creates an AI voice that sounds exactly like the recording. Get a Voice ID in seconds.
3. Write Script: Add script lines, assign each line to a specific voice, and type what that person says. Reorder lines, use templates, or import from a file.
4. Generate Audio: Click "Generate Audio" and VoxKit sends each line to the AI provider with the assigned voice. Audio clips are returned and played sequentially.
5. Play & Download: Play all clips in sequence, or download individual clips as MP3 files.

4 Voice Cloning Providers (BYOK)

VoxKit supports 4 AI voice providers plus a free browser fallback. Add your API key from any provider — keys are stored in your browser only, never on our server.

Browser TTS (Free Fallback)

No API key? No problem. VoxKit includes a free Browser TTS mode that uses your device's built-in text-to-speech engine. Quality varies by browser/device, but it works on Chrome, Edge, Safari, and most mobile browsers. Perfect for testing the tool before getting a provider key.

Voice Cloning — How It Works

Voice cloning creates an AI voice that sounds exactly like a specific person. Here's the process:

1. Record or Upload: Record 10-60 seconds of someone speaking (clear audio, no background noise). Or upload an existing audio file (MP3, WAV, M4A).
2. Clone: Click "Clone Voice" — VoxKit sends the sample to ElevenLabs, which analyzes the voice characteristics (pitch, tone, accent, pacing, breathing patterns).
3. Get Voice ID: ElevenLabs returns a unique Voice ID. This ID represents the cloned voice in all future generations.
4. Generate: Type any text, select the cloned voice, and the AI speaks in that person's voice — indistinguishable from the original.

The cloned voice can say anything, in 29 languages, with adjustable stability (conservative vs. expressive), similarity (how close to original), and style (how stylized).

Script Editor Features

Per-Voice Settings

Each voice has independent settings that control how the AI speaks:

• Stability (0-100%): Low = more expressive and varied. High = more consistent and conservative.
• Similarity Boost (0-100%): How closely the AI voice matches the original sample. Higher = closer match.
• Style Exaggeration (0-100%): How much the AI exaggerates the voice's style. Higher = more stylized.
• Speed (0.5x - 2x): Playback speed. 1x = normal, 0.5x = slow, 2x = fast.
• Language: 29 languages for multilingual TTS.

Audio Enhancement

VoxKit applies enhancement algorithms to improve audio quality:

Mobile + PC Compatibility

VoxKit works on both mobile and desktop:

• Microphone Recording: Uses navigator.mediaDevices.getUserMedia() — works on mobile Chrome, Safari, and desktop browsers
• Audio Playback: HTML5 <audio> elements — works on every device
• Responsive Design: Mobile-first layout with touch-friendly controls
• Browser TTS: Uses device's built-in speech synthesis — works offline on most mobile devices

Use Cases

Auto-Update System (Deploy Once, Updates Forever)

VoxKit uses an auto-update architecture — you deploy the HTML block ONCE, and all future features are automatically loaded from the server. No need to re-paste the HTML block when new features are added.

How It Works

1. The HTML block contains a bootstrap loader that runs on every page load.
2. The loader checks the server for the latest voxkit-remote-features.js file via the get_features_manifest and get_remote_features API endpoints.
3. If a newer version exists, the new JavaScript code is automatically injected and executed.
4. The version is cached in localStorage for instant loading on subsequent visits.
5. The check runs on page load, then every 24 hours automatically.

What This Means For You

• Deploy once: Paste the HTML block into WordPress one time. That's it.
• Automatic updates: When the developer adds new features to voxkit-remote-features.js on the server, all deployed HTML blocks auto-load them within 24 hours (or on next page load).
• No re-deployment: You never need to re-paste the HTML block again.
• Manual check: Click "Check for Updates" in the Advanced section to fetch updates immediately.
• What's New modal: Click "What's New" to see the changelog and current version.
• Version display: Footer shows current auto-update version with a green "Live" indicator.
• Update notifications: "VoxKit updated!" badge appears when new features are loaded.

Currently Auto-Loaded Features

The voxkit-remote-features.js file (v1.0.0) currently auto-loads 15 features:

• Auto-update notification badge — shows when new features load
• Version display in footer — current version + live indicator
• "What's New" button + modal — full changelog
• "Check for Updates" button — manual update check
• Live system status indicator — provider, voices, version
• Auto-backup every 5 minutes — last 10 backups saved
• Restore from auto-backup — recover unsaved work
• Random tip rotation — helpful hints on load
• Editing indicator — "● Editing..." when typing
• Crash recovery prompt — restore work after browser crash
• Performance monitor — logs slow requests to console
• Welcome back message — greets returning users
• Session time tracking — total time in VoxKit
• Feature usage tracking — anonymous, local only
• First-visit demo — shows welcome tip for new users

For Developers: Adding New Features

1. Edit voxkit-remote-features.js on the server — add your new feature code (use window.VoxKitAPI, window.VoxKitState, window.VoxKitAddToolbarButton, window.VoxKitCreateModal).
2. Bump the version in voxkit-features-manifest.json (e.g. "1.0.0" → "1.1.0").
3. Add the feature name to the features array in the manifest.
4. Add a changelog entry in the manifest.
5. Save both files — all deployed HTML blocks auto-update within 24 hours.

That's the entire workflow. No HTML editing, no re-deployment, no user action needed.

Auto-Updateable AI (Like Plagiarism Checker)

All AI features use a 4-key fallback system (Gemini → GLM → Groq2 → Groq1) — exactly like the Plagiarism Checker. No user API key needed for any AI feature. The system auto-routes through whichever server-side key is working, with 2 retries per provider.

Auto-Update Mechanisms

• License Auto-Refresh — Premium license status is refreshed every 5 minutes. Auto-detects when license is exhausted and switches to Free tier.
• Voice Catalog Auto-Update — Azure and Google voice catalogs are cached locally for 24 hours, then auto-refreshed from the server.
• Music/SFX Library Auto-Update — Background music and SFX libraries auto-refresh every 24 hours.
• Version Check — VoxKit checks for new versions and features on load, then daily. Shows update badge when new features are available.
• Provider Health Monitor — Provider endpoint reachability checked every 10 minutes. ElevenLabs health verified live.
• Auto-Save — Script and voices auto-save every 30 seconds with visual indicator showing last saved time.

AI-Powered Features (11 New — Server-side, no API key needed)

• AI Translate Script — Translate entire script to 13 languages (Spanish, French, German, Italian, Portuguese, Chinese, Japanese, Korean, Hindi, Arabic, Russian, Dutch, English). Preserves speaker names. One click applies translation + updates voice language setting.
• AI Show Notes Generator — Creates podcast show notes from your script: 2-3 sentence summary, key points bulleted, timestamps with topics, tags, and call-to-action. Copy to clipboard for podcast hosts.
• AI Social Media Posts — Generate platform-optimized posts for Twitter/X, LinkedIn, Facebook, and Instagram. Includes hashtags. One-click copy per post.
• AI Episode Titles — Generate 5 SEO-friendly, catchy episode titles from your script content. Mix of styles: question, list, bold claim, mystery, how-to. Click to set as episode title.
• AI Voice Casting — AI acts as casting director, recommending which voice should say each line based on tone, emotion, and content. Includes reasoning and suggested emotion per line. One click applies all recommendations.
• AI Long Text → Script — Paste any long-form text (article, blog post, transcript up to 8000 chars) and AI converts it into a multi-voice conversational script. Choose 2-4 speakers and 6-16 lines.
• AI Chapter Auto-Tagger — AI analyzes your script and identifies topic shifts where chapter breaks should occur. Inserts [CHAPTER: Title] markers automatically.
• AI Pronunciation Suggestions — AI identifies difficult-to-pronounce words (technical terms, names, foreign words) in your script and suggests phonetic respellings. One click adds all to your Pronunciation Dictionary.
• AI Emotion Matching — AI recommends the best emotion tag (happy, sad, angry, calm, whisper, excited, serious, neutral) for each line based on content. Includes intensity score. One click applies all.
• AI Smart Auto-Complete — "AI Next" button on each script line. AI suggests the next line of dialogue based on the conversation context.
• Voice Quality Scorer — When you upload or record a voice sample, AI automatically analyzes it and gives a quality score (0-100) with recommendations. Badge displayed on each voice card.

Additional Advanced Features (14 New)

• VU Meter During Recording — Real-time audio level meter (green→orange→red) shown during microphone recording. Helps you maintain consistent recording levels.
• Audio Spectrum Analyzer — Real-time frequency spectrum visualization during audio playback. 32-band FFT display.
• Live Caption Preview — When audio plays, the current line's text appears as a floating caption at the bottom of the screen. Great for accessibility.
• Episode Auto-Numbering — Remembers your last episode and season number. Auto-suggests the next episode (S1E1, S1E2, ...). One click sets it as current. Reset counter anytime.
• Cross-Episode Voice Memory — Save your voice configurations as named sets. Load them in future episodes to maintain consistent voices across your podcast series. Save up to 20 voice sets.
• Auto-Trim Silence — Automatically detects and removes silence at the beginning and end of voice recordings. One click per voice, or enable auto-trim in preferences.
• Auto-Normalize Volume — Peak-normalizes voice recordings to -3dB. Ensures consistent volume across all voices. Shows the gain applied in dB.
• Episode Trailer Generator — Automatically picks the 4 most spread-out clips from your generated audio and combines them into a 15-second trailer preview. Download as single WAV.
• Script Difficulty Analysis — Analyzes reading level of each line (easy/medium/hard) based on word count, sentence length, and long-word ratio. Helps identify lines that may be hard for TTS to pronounce clearly.
• Audio Watermarking — Adds a barely-audible (15kHz at -40dB) marker to the end of generated audio. Detectable but inaudible — useful for tracking unauthorized use.
• Recording Quality Presets — Choose recording quality: Studio (48kHz/256kbps), Podcast (44.1kHz/128kbps), Phone (16kHz/64kbps), Voice Memo (22kHz/96kbps).
• Persistent Preferences — Auto-save, auto-translate, auto-score-quality, auto-trim-silence, and watermark settings saved across sessions.
• Auto-Retry Failed Generations — When a line fails to generate, the system automatically retries with browser TTS as fallback.
• Smart Auto-Save Indicator — Floating "Saved [time]" indicator appears top-right whenever the script or voices are saved. Persists for 2 seconds.

Pro Features (40+ Advanced Tools)

All features below are unlocked in VoxKit. Click the corresponding button in the Advanced section of the tool to use them.

AI-Powered Writing & Analysis

• AI Script Writer — Generate a complete podcast, interview, debate, audiobook, news, comedy, educational, or storytelling script from a topic. Powered by GLM-4. Choose 2-4 speakers, 6-16 lines, 6 languages.
• Sentiment & Mood Analysis — AI analyzes each line's emotion (happy/sad/angry/calm/excited/serious/whisper/neutral), sentiment score (-1 to +1), and reading level (easy/medium/hard). One click auto-applies the suggested emotion per line.
• Spell Check — GLM-powered grammar and spelling check across your entire script. Lists errors with suggestions.
• Voice Mood Matching — AI suggests the best emotion tag for each line based on its content.

Audio Production

• Audio Effects Chain — EQ presets (Podcast, Audiobook, Broadcast, Warm, Bright, Phone), Reverb (Small Room, Hall, Stadium, Cave), Compression (Light/Medium/Heavy), LUFS Loudness Normalization (-23, -16, -14, -19 LUFS), Noise Gate threshold.
• Post-Processing — Adjust overall speed (0.5x-2x), pitch (-12 to +12 semitones), and volume (-12 to +12 dB) of generated audio.
• Multi-Track Export — Export each voice as a separate WAV stem. Perfect for mixing in DAWs like Audacity, Reaper, or Pro Tools.
• Single WAV Export — Combine all generated clips into one continuous WAV file with proper encoding.
• MP3 Bitrate Selection — 64 / 128 / 192 / 320 kbps for size vs. quality tradeoff.
• Real-time Waveform Visualization — Live audio waveform during microphone recording.
• Audio Trimming & Noise Reduction — Cut silence from beginning/end, apply noise gate filter.

Voice Management

• Voice Library Browser — Browse 400+ Azure voices and 380+ Google voices by name, language, gender, or style. Searchable catalog with one-click apply.
• SSML Editor — Edit Speech Synthesis Markup Language for fine control. Templates for <break>, <prosody>, <emphasis>, <phoneme>, <say-as>, <sub>.
• A/B Voice Comparison — Generate the same text with two different voices side-by-side. Play both to pick the best.
• Voice Avatars — Upload an image for each voice (auto-resized to 100x100). Visual identification in long projects.
• Voice Cloning History — Track all cloned voices with original Voice IDs, provider, and timestamp.

Publishing & Distribution

• Subtitle Generation — Generate SRT (SubRip) or VTT (WebVTT) subtitles synced to your script timing. Download as a file.
• RSS Feed Generator — Create a podcast RSS XML feed with iTunes tags, season/episode numbers, categories, cover art URL, audio enclosure.
• Episode Metadata — Title, description, author, season, episode, language, categories, cover art URL, audio URL, duration. Saved locally and reusable.
• Cover Art Generator — Generate 1400×1400 podcast cover with title, subtitle, color gradient (linear/solid/diagonal/radial), adjustable font size. Meets Apple Podcasts spec.

Script Management

• Project Templates — Save the current voices + script as a reusable template. Load or delete saved templates anytime.
• Script Import — Import from .txt, .srt, .vtt, or .json. Auto-detects "Name: text" format for plain text.
• Script Export — Export as TXT (with speaker labels), JSON (full project), or SRT (subtitle format).
• Find & Replace — Search across all script lines. Replace one occurrence or all at once.
• Multi-Select Lines — Checkbox-select multiple lines for bulk operations: delete, duplicate, change voice, bookmark.
• Bookmarks — Mark important lines with a star. Quickly find them later.
• Collaborative Notes — Add a per-line note for team collaboration. Notes are saved with the project.
• Line Color Coding — Each voice gets a unique color dot. Quick visual identification of who speaks where.

Music & Sound Effects

• Music Library — 5 built-in royalty-free background tracks (Ambient Calm, Upbeat Corporate, Tech Pulse, Soft Piano, Corporate Warm) from Pixabay.
• SFX Library — 8 sound effects (Applause, Laughter, Whoosh, Ding, Phone Ring, Door Close, Rain, Typewriter). One-click set as transition sound.

Visualization & Stats

• Speaker Timeline View — Visual timeline showing when each speaker talks, with color-coded segments and a time ruler.
• Statistics Dashboard — Total generations, total lines, total audio minutes, voice count, script line count, pron entries, breakdown by provider, recent generation history.
• Audio Preview Cache — Cache up to 20 generated audio clips locally. Re-generating the same line is instant.
• Generation History — Last 5 generations with timestamp, line count, duration, and provider.

UX & Accessibility

• Onboarding Tour — 5-step interactive walkthrough for first-time users. Auto-shows on first visit. Replayable anytime.
• Hotkeys Reference Card — Press ? anytime to see all keyboard shortcuts (Ctrl+Enter/S/D/F/B/Shift+N).
• PWA Installable — Install VoxKit as a standalone app on desktop or mobile. Works offline for browser TTS.
• Dark Mode — Toggle dark theme. Persists across sessions.
• Auto-Save — Script and voices auto-save every 30 seconds. Visual indicator shows when saving.

Pricing

Frequently Asked Questions

Ready to Create Your Podcast?

Clone voices, write scripts, and generate multi-voice audio. Free: 2/day. Premium: 10 for $5.