PeaBox AI Speech Studio

Overview

PeaBox AI Speech Studio is a local-first desktop studio that combines text-to-speech, transcription, subtitles, audiobook generation and voice-provider management in one interface.

All processing runs on your machine. When cloud voices are used, only the request payload reaches the provider. The studio does not upload your files for analysis or persist them.

Features

• Text-to-speech with local and cloud voices
• Transcription from uploaded audio/video files
• Generate subtitles and transcripts
• Audiobook pipeline with chapter segmentation
• Import, export, merge, split and trim audio files
• Multiple cloud voice providers: Azure Speech, Amazon Polly, Google Cloud TTS
• Local synthesis using pyttsx3 on supported platforms
• Batch jobs and voice selection UI
• History of generations and downloads

Providers

Local: pyttsx3 using system-installed speech engines. Cloud: Azure Speech (Text-to-Speech REST), Amazon Polly (SynthesizeSpeech), Google Cloud Text-to-Speech (SynthesizeLongText).

Cloud Voice Configuration

Each cloud provider needs an API key configured in the app settings. The app never persists credentials in plaintext; they are stored in OS keyring.

When a cloud provider is selected, the request is sent directly to the provider's endpoint. Audio responses are written to local files; no analysis results are uploaded to PeaBox.

SSML Support

Supports Speech Synthesis Markup Language tags for prosody, emphasis, breaks, silence, pronunciation and sentence boundary markers. Use plain text mode for simple generation or SSML mode for precise control.

Audiobooks

Split chapters into segments with configurable pause durations, paragraph pauses and emphasis control. Generate one MP3 per chapter or merge into a single file.

Subtitles & Transcripts

Transcribes audio/video and generates subtitle files (SRT/VTT) and plain-text transcripts. The transcript is editable and can be exported alongside the audio.

Architecture

PeaBox AI Speech Studio
"œ"" app.py               Main Streamlit application
"œ"" providers/           Cloud provider adapters
"‚   "œ"" azure.py
"‚   "œ"" polly.py
"‚   """" google.py
"œ"" local_synthesis/     Local voice engine wrapper
"œ"" audio_processing/    Clip, trim, merge, split helpers
"œ"" export/              MP3, WAV, SRT, VTT, TXT writers
"œ"" tests/               Unit and integration tests
"""" requirements.txt

Requirements

• Python 3.9+
• Streamlit (desktop app runtime)
• pyttsx3 for local synthesis
• Azure Speech, Amazon Polly or Google Cloud TTS credentials for cloud voices
• ffmpeg for audio merging and format conversion (installed locally)

Roadmap

Add waveform editor, more SSML templates, batch download queue, cross-platform TTS, and local whisper-based transcription.