Create
Set creative direction and composition inputs, then choose an available rendering path.
Creative intelligence
Callio
AI music built around the creator.
Working prototype — active developmentRequest beta access
Why Callio exists
AI music built around the creator.
Callio begins with explicit creative direction and a deterministic symbolic song plan before audio rendering.
Creator-editable, instrument-pure stems are the objective. Current full-song neural paths primarily produce mixes that require separation, so separated stems are not described as native stems.
Desktop workflow
From idea to a reviewable project.
The current application is a Tauri/React desktop prototype connected to a modular inference and rendering service.
Studio
Review generated material, work with mixing controls, and keep changes inside the project.
Library and history
Return to prior projects and generated work without presenting an ephemeral one-shot interface.
Mix and export
Prepare selected outputs for export while keeping native symbolic material distinct from separated neural mixes.
Creator-control principles
Direction, choice, review, and consent stay visible.
Direct the composition
The implemented composition and theory core keeps the symbolic plan inspectable before rendering.
Choose the rendering path
Selectable local and controlled cloud routes expose different quality, privacy, and workflow tradeoffs.
Review before export
Mixing, project history, and export keep the creator inside the final decision loop.
Route voices explicitly
Consent references and destination controls are implemented in current voice-routing UI/API paths.
Workflow architecture
Structure first; rendering second.
This diagram describes the current public architecture and does not imply universal renderer quality.
Design / architecture material
Creator-directed music workflow
Callio separates musical planning from audio realization.
- Creative direction
The creator supplies musical, lyrical, structural, and performance intent.
- Composition and theory
The deterministic core establishes an inspectable symbolic song plan.
- Renderer-specific planning
Performance planning and the experimental Prose Engine support selected engineering paths.
- Renderer orchestration
Selectable local and controlled cloud paths interpret the plan with explicit fallbacks.
- Mixing and stems
Current outputs move through mixing, export, and—where necessary—estimated separation.
- Creator review
The creator reviews, revises, and decides what leaves the system.
Scope: The proposed standalone Arrangement Engine is not built. The Prose Engine is implemented as an experimental module, but normal desktop-path integration is not yet established.
Rendering paths
Different engines produce different kinds of control.
Location, output structure, and evidence state stay attached to each path.
| Path | Execution | Output | Current state |
|---|---|---|---|
| Symbolic renderers | Local | Structured audio plus editable symbolic material | Implemented with HQ, fast/FluidSynth, and NumPy fallback paths |
| MiniMax Music 3 | Local or controlled hosted path | Full-song mix; separation may be used afterward | Integrated; not a native-stem renderer |
| ACE-Step | Optional configured worker | Full-song and voice-conditioned paths | Staging integration established; broad quality is not claimed |
| Voice routing | Path-specific local or cloud destination | Voice-conditioned render inputs | Consent reference and destination gates exist in current UI/API paths |
| Prose Engine | Experimental engineering module | Song-specific prose/performance planning | Implemented and tested; normal desktop flow not established |
Current capabilities
A working desktop prototype under active development.
Callio creates a deterministic symbolic song plan through its composition and theory core before rendering audio.
- Scope
- Current composition/theory implementation and renderer-specific performance planning.
Limits
- The proposed standalone Arrangement Engine is not built.
- Neural renderers can reinterpret the plan.
- Not every renderer produces native stems.
The current desktop application is a working prototype under active development.
- Scope
- Current desktop branch and local application architecture.
Limits
- The repository does not contain a current public screenshot or release-ready media set.
Callio's product objective is creator-editable, instrument-pure stems; current full-song neural renderers primarily produce mixes that require separation.
- Scope
- Current MiniMax and ACE-Step full-song paths.
Limits
- Separated stems must not be described as native instrument-pure stems.
Real application
Current product capture is being prepared.
Application screenshots
Current application screenshots and screen recordings are not yet published. No mock interfaces are shown as product captures.
Media and generation evidence
Product media is being prepared.
Screenshots, demo audio, and generation records will be added as they become available for public release.
Current screenshots
Create, Studio, Library, renderer-selection, mixing/export, and voice-consent screenshots are not yet available.
Demo audio
Demo audio is not currently published.
Generation records
Renderer and vocal-routing experiments are not yet published as evidence of general quality.
Voice and consent
Permission and destination are explicit system inputs.
Current status and limits
Working features, product objectives, and experiments remain distinct.
Current boundary
Callio is a working prototype in active development. Current product screenshots and demo audio are not yet published, and renderer behavior differs by path.
- The proposed standalone Arrangement Engine is not built.
- Neural renderers can reinterpret the plan.
- Not every renderer produces native stems.
- The repository does not contain a current public screenshot or release-ready media set.
- Separated stems must not be described as native instrument-pure stems.
In development
Roadmap—not demonstrated capability.
Standalone Arrangement Engine
Renderer qualification
Native-stem strategy
Broader genre validation
Workflow refinement
Callio beta
Help shape a creator-controlled music workflow.
Request beta access or product updates directly from the Callio team.
Data-minimalCreator-focusedNo automatic enrollment
Tell us how you make music and what you would like to explore with Callio.
Request beta accessOpens your email application. Nothing is submitted through this website.Work with Korelis Labs
Evaluate a creator-controlled workflow.
Creators, musicians, evaluators, and technical collaborators can start a focused conversation.