Echova Documentation

Overview

Echova Studio is organized around the task you want to finish. Most workflows start with one of three things: a voice, a script, or a media file.

Use these docs as practical workflow guides. Each article explains what to prepare, which controls matter, how to review the result, and what to change when the output needs work.

Choose your path

• Voices: choose a built-in voice, save favorites, clone a permitted speaker, design a new AI voice, or add reusable emotion styles.
• Studio: turn text into speech, choose the voice and model, tune generation settings, and create final voiceover audio.
• Audio Dubbing: translate or revoice an audio file while keeping the job easy to review.
• Video Dubbing: dub a video file and check timing before exporting the result.
• Voice Isolator: separate speech from background audio before dubbing, transcription, or cleanup.
• Transcribe: create, review, edit, and export transcripts from audio or video.
• Audio Library: find, preview, compare, export, reopen, or delete generated audio files.
• Models: install local models, check dependencies, and understand which model is used for each workflow.
• Settings: manage storage, API and MCP endpoints, license details, updates, and support information.
• Support: open support resources and collect the details needed before asking for help.

First setup

Before a long generation job, check the basics:

• Install the models needed for your workflow in Models.
• Confirm your storage location in Settings, especially if you work with large audio or video files.
• Create, save, or choose a voice in Voice Lab before opening Studio.
• Keep a short test script ready so you can check a voice before generating a full project.

Recommended order

If you are new to Echova, start with this path:

1. Read Voice Lab to understand where voices live.
2. Try Create an AI Voice or Clone Voice, depending on whether you have a source recording.
3. Generate a short sample in Studio.
4. Review and compare the output in Audio Library.
5. Move to dubbing, transcription, or voice isolation when you have a source file that needs processing.

When something looks wrong

If a page does not show the control you expect, make sure you are on the matching app tab and that the required model is installed. If a generated file sounds wrong, open the latest job from the library and check the voice, model, language, emotion, speed, and advanced settings used for that run before starting over.