TTSForge

TTSForge is an audiobook-focused command-line frontend for Readio. It gives EPUB audiobook workflows a concise CLI; Readio owns the persistent project, speech planning, synthesis, composition, progress, and export.

TTSForge does not implement or bundle a parallel speech engine. Use Readio's public application API and its engine integrations; TTSForge maps audiobook choices to those services.

TTSForge requires Readio >=0.3.4, which provides the public audiobook API, expanded synthesis-resolution contract, and persisted project settings used by guided setup. Install TTSForge normally, then add a Readio engine integration as needed; see Installation for setup and development instructions.

What it does

• Inspect EPUB metadata and chapter structure.
• Create or reuse a persistent Readio project for an audiobook.
• Select chapters, preview speech, plan a project, and check authoritative project status.
• Export M4B audiobooks or formats supported by Readio's generic export service.
• Discover voices, models, engines, and formats through Readio.
• Use Readio's configuration and SSMD authoring/checking services.

Quick start

Install TTSForge and a Readio engine integration using the Readio extra for your platform; see Installation for details. Then run:

ttsforge convert novel.epub

In an interactive terminal, a new project shows the detected chapters and prompts for a selection, then guides you through synthesis settings not supplied by CLI or already saved in the project. Readio's model and voice choices appear only when a choice is needed. The resolved setup is saved to the Readio project before confirmation or build. Later runs reuse those values, so a failed build can be retried without repeating setup prompts. Use --reconfigure to review saved choices with their current values as defaults; explicit CLI options pin and update their corresponding settings. Build progress updates live by chapter.

For automation, specify --chapters and any synthesis options that should override the project. --yes skips only final confirmation; --non-interactive and --json never prompt and reuse saved setup when available. Without an explicit chapter selection, non-interactive runs include all chapters for a new project. Useful follow-up commands:

ttsforge list novel.epub
ttsforge convert novel.epub --chapters 5-17 --fresh
ttsforge status novel.readio
ttsforge preview novel.epub
ttsforge voices --language en-us
ttsforge doctor

By default, the project is novel.readio beside novel.epub; the default output is novel.m4b. Re-running the command reuses that Readio project's chapter scope and saved synthesis/export settings. See Projects and outputs for reuse, fresh projects, and legacy workspaces.

Readio engines and configuration

Engine availability and voices depend on Readio's installed optional engine integrations and local configuration. Discover the active catalog and diagnose the environment with:

ttsforge engines
ttsforge models
ttsforge voices --engine kokoro
ttsforge formats
ttsforge doctor

TTSForge reads and updates Readio configuration rather than keeping a separate TTSForge backend configuration:

ttsforge config show
ttsforge config set reader.voice af_heart
ttsforge config set reader.speed 1.1

Migration notes

This is a breaking change from the former TTSForge-owned Kokoro conversion pipeline. Old renderer workspaces are not Readio projects and cannot be resumed by this frontend. They are left untouched; use --fresh or an explicit new --project path to start separately. Backend-specific commands such as read, sample, demo, download, and phonemes are no longer part of TTSForge. See the migration guide before switching an existing setup.

Documentation

• Installation
• Quick start
• CLI reference
• Project and output lifecycle
• Readio configuration
• Voices and discovery
• SSMD tools
• Migration guide
• Python API boundary
• Command-line examples

Implementation details and the public service contracts are documented in Readio's API guide and project guide.