ColvriView project
Planned capability

Shape a local voice workflow.

Generate and compare audio while source text and results remain on your machine.

Voice workspace Ready locally
Script

A warm narration about rain moving through the Andean cloud forest.

Rain over the valleyNatural voice · en-US
00:18

Colvri is designed to give voice models a focused workspace without folding every runtime and dependency into the core application.

A focused interface

Write a script, choose an installed voice, compare results, and keep useful takes with the project that produced them.

Local source material

The intended workflow keeps prompts, reference material, generated audio, and project metadata in the user-controlled environment.

An optional module

Voice support is planned as a capability you can add independently, with its own adapter and isolated runtime requirements.

From a private script to a take worth keeping.

The voice flow is organized around comparison, not a one-shot generation button.

Prepare the script

Keep the source text and project context together before selecting an installed voice model.

Compare local takes

Generate alternatives with visible voice, language, and runtime choices instead of hiding the configuration.

Keep the useful result

Attach the selected audio and its settings to the project that produced it.

A listening workspace, not an anonymous output tray.

Scripts, takes, timing, and the active voice remain visible in one review surface so a useful result keeps its context.

Planned local workspace
Voice workspaceProject saved locally
Source script

A calm narration about rain moving across the Andean cloud forest.

Cloud forest story · draft saved
Selected take00:18
Natural voice · English (US) · local preview

Know what the voice module brings with it.

Before installation, the adapter should explain the model source, language coverage, local runtime, storage footprint, and license.

Decision before downloadIllustrative interface · exact values belong to each validated adapter.
SourceExact model artifactNo ambiguous latest reference
RuntimeIsolated voice workerKept outside the Colvri core
StorageDeclared before downloadModel-specific, never invented
TermsVoice model licenseRead independently from Apache-2.0

Voice can begin the story without owning the whole studio.

Add image or video later and keep the narration connected to the same project. Each capability stays optional and replaceable.

Return to the product overview

Voice questions, with the boundaries kept visible.

These answers describe the intended capability workflow while keeping pre-release status, adapter requirements, and third-party terms explicit.

The planned workflow runs generation locally and keeps source text in the user-controlled project environment; it does not require a proprietary cloud upload.