Voicebarn turns any script into clean, natural narration — 100% on your machine, powered by the Piper neural TTS engine. It's for YouTubers, course creators, audiobook hobbyists, and indie devs who are tired of two things: per-character subscription caps, and uploading unpublished scripts to someone else's cloud. ElevenLabs Creator is $22/mo (~$264/yr) with a 100,000-character monthly cap; Voicebarn is $34 once with no cap at all — it pays for itself in about two months. Honest framing: these are clean neural narration voices, not celebrity voice cloning.

• Local neural TTS via Piper — fully offline after the one-time engine + voice download

• 8 curated starter voices: English US/UK, German, Spanish, French

• Per-paragraph voice and speed control (0.5x–2.0x), inheriting document defaults

• Inline pause tags (under pause 500msover ) — sample-rate-matched silence, honestly scoped SSML-lite

• Instant per-paragraph preview, cached by content hash

• WAV or MP3 export with a bitrate picker (128–320 kbps) via bundled ffmpeg

• Batch mode: folder of .txt scripts in, folder of narrated audio files out

• Local project save/load — no account, no telemetry, text never leaves your PC

Runs as a Windows desktop app.

What's included:

• Windows installer download (build via npm run dist)

• Full source access — https://github.com/bensblueprints/voicebarn (MIT)

• Lifetime v1 updates

• Setup guide (README quick start)

• Premium tiers coming later