🗣️ Voicebarn
Unlimited natural-sounding text-to-speech on your own machine.
ElevenLabs Creator costs $22/month — $264 a year — and still caps you at 100,000 characters a month. Voicebarn runs Piper, a fast local neural TTS engine, on your own machine: unlimited characters, per-paragraph voice and speed control, batch folder conversion, and your unreleased scripts never touch anyone's cloud.
Voicebarn screenshot is being captured — the app is shipped and real.
Features
Local neural TTS via Piper
Clean, natural narration voices, 100% offline after the one-time engine + voice download.
Per-paragraph control
Every paragraph gets its own voice + speed override (0.5x–2.0x), inheriting document defaults when left blank.
Inline pause tags
Drop
Instant preview
Synthesize and play any single paragraph before you export, with results cached by content hash.
WAV or MP3 export
MP3 via bundled ffmpeg with a bitrate picker (128/192/256/320 kbps).
Batch mode
Point it at a folder of .txt files and get one narrated audio file per input — chapter-by-chapter audiobook workflow.
8 curated starter voices
English (US/UK), German, Spanish, French — medium-quality Piper voices, each downloaded once from Hugging Face.
100% private
No telemetry, no account, no network calls except the clearly-surfaced one-time engine and voice downloads.
Voicebarn vs ElevenLabs (Creator)
ElevenLabs (Creator) at $22/mo runs roughly $264/year — $528 over two years. Voicebarn is $34, once.
| Voicebarn | ElevenLabs (Creator) | |
|---|---|---|
| Price | $34 once | $22/mo, forever (~$264/yr) |
| Characters | Unlimited | 100,000/mo cap (Creator) |
| Privacy | Text never leaves your PC | Uploaded to their cloud |
| Works offline | Yes | No |
| Per-paragraph voice/speed | Yes | No |
| Batch folder processing | Yes | Limited/paid tier |
| Voice cloning | No — narration voices only | Yes |
| Account required | No | Yes |
Pays for itself in under 2 months versus ElevenLabs Creator. Year two costs you $0; theirs costs another $264.
Three steps, no subscription
Buy once on Whop
One-time $34 for the packaged Windows installer with lifetime updates — no terminal, no build step.
First run downloads the engine
The Piper engine (~22 MB) from official GitHub releases and your default voice (~60 MB) from Hugging Face, with a visible progress bar. After that it's fully offline.
Write, preview, export
Assign voices and speeds per paragraph, drop pause tags where you need a beat, preview any paragraph, export WAV or MP3 — or batch a whole folder.
Honest answers
Is it really free on GitHub?
Yes — the full source is MIT at github.com/bensblueprints/voicebarn, always. $34 buys the signed installer, 1-click setup and updates instead of npm i && npm start.
Can it clone my voice or sound like a specific person?
No — and ElevenLabs is genuinely better if that's what you need. Voicebarn ships clean neural narration voices, great for videos, courses, audiobooks, IVR prompts and accessibility — not celebrity impressions or voice cloning. That's a different product category.
Does it support SSML?
Only an SSML-lite pause tag:
How does the quality compare to ElevenLabs?
Piper's voices are natural, clean narration voices — very good for the job, but ElevenLabs' newest cloud models are more expressive. What you get in exchange: unlimited characters, total privacy, offline operation and no bill ever again.
Does my text get uploaded anywhere?
Never. Synthesis runs locally via Piper (ONNX runtime on your CPU). The only network calls are the one-time engine and voice downloads, both clearly shown in the UI.
Own Voicebarn forever
$34 once. Signed installer, 1-click setup, updates included. No renewal, no account with us, no meter. Or build it yourself from the MIT source — it's the same app.