Narrator-TTS

AI-Powered Narration with Custom Voices

Narrator-TTS is a custom local pipeline producing natural, expressive narration for long-form content — audiobooks, scripture, essays, and storytelling. These samples showcase original voices built from the ground up.

How These Were Made

01

Custom Voice Profiles

Original voices built from reference audio — Dazlarus (warm, narrative male), Tigress (expressive, dynamic female), and Pastor Jeff (measured, pastoral) — each fine-tuned for their intended register.

02

Intelligent Text Segmentation

Source texts are automatically segmented into narration-sized chunks at natural boundaries — paragraph breaks, chapter divisions, and semantic stopping points.

03

Generation with QA

Each segment is synthesized and then verified against the source text using ASR transcription — catching dropped words, garbled audio, and pronunciation errors before assembly.

04

Automated Assembly

Narrator-TTS manages the entire pipeline — multi-worker parallel generation, retry logic, VRAM management, and final assembly with crossfaded breath spaces into a polished audio file.

Voice Comparison — Gettysburg Address

The same text narrated by two voices. Notice how each brings a different character and emotional register to Lincoln's words.

Dazlarus

The Gettysburg Address

Warm, measured male narration. A grounded storytelling voice suited for long-form fiction and narrative nonfiction.

Voice: Dazlarus Method: XTTS v2
Tigress

The Gettysburg Address

Expressive, dynamic female narration. A versatile voice with strong emotional range — equally suited for scripture, drama, and literary narration.

Voice: Tigress Method: XTTS v2

Scripture Narration — Colossians (NIV)

Full chapter narration demonstrating the pipeline's handling of verse structure and devotional tone across a sustained reading.

Custom Production Available

Need professional narration for your book, article, scripture series, or creative project? Custom voice development, multi-voice productions, and full pipeline assembly are available.

Voices can be tailored to your material. Turnaround depends on length and complexity. Rate sheets and sample packages available on request.