Systems
Compatible audio systems
Make audio projects, interfaces, state, and evidence portable across qualified browser, native, DAW, and hardware contexts.
Syntho Research
Partners welcomeBuilding the future of music.
We research how music can be generated, represented, evaluated, visualized, and turned into useful tools. Our work lives where papers, models, creative practice, and production systems meet—and grows through collaboration with researchers, audio companies, artists, and open communities.
Research program
Sound, structure, intelligence, and the tools between them.
The Engine is one important foundation, but it is not the whole program. The broader Syntho toolkit connects computational music theory, models, datasets, visualization, musical agents, and a growing plugin ecosystem so research and creative tools can work together.
Sound
Generative sound and coherent kits
Generate useful collections and instruments whose sounds work together—not merely convincing samples in isolation.
Structure
MIDI and symbolic music
Develop models, tools, and representations that preserve timing, harmony, rhythm, editability, and musical control.
Evidence
Musical evaluation
Build benchmarks, listening studies, fixtures, and interaction studies that test musical usefulness rather than one metric alone.
Understanding
Theory and model visualization
Make harmony, rhythm, tension, structure, and model decisions easier to perceive, inspect, and direct.
Interaction
Musical agents
Study how musicians can guide intelligent tools, hear and inspect changes, recover work, and remain the authors of the result.
Current and future work
Questions already becoming projects.
Some directions already have code, tools, models, or papers behind them; others are interests we want to open with the right partners. Félix has collaborated with researchers from the Metacreation Lab for Creative AI at Simon Fraser University to help bring selected research projects to the web. We keep active work, research evidence, and future intent distinct while showing how they connect.
Engine and compatible audio systems
The question
Researchers can prove an algorithm yet still face a second engineering project to add controls, saved state, interfaces, realtime safety, host integration, and DAW delivery.
Current direction
Develop the compiler, SDK, DSP, UI kit, runtime contracts, and qualification methods as one foundational project that other research can use.
Coherent drum-kit generation
The question
Generating one plausible drum hit does not establish that a model can create a role-complete kit whose sounds belong together and remain musically useful.
Current direction
Study structured kit generation, set-level coherence, role-aware evaluation benchmarks, rights-clean data, and playable instrument outcomes.
MIDI and symbolic music
The question
Musical models need representations and controls that preserve timing, structure, theory, editing, and live interaction—not just plausible token sequences.
Current direction
Develop MIDI generation tools and models, symbolic-music data techniques, constraints, visualizations, and agent interactions in working musical contexts.
Music theory made visible
The question
Harmony, rhythm, tension, and structure are rich relationships, but most software hides them behind notation or opaque model behavior.
Current direction
Explore visual and interactive strategies that make musical theory, model choices, and generated structure easier to perceive and direct.
Work with us
Bring a question, a contribution, or a useful constraint.
- You have a hard music, audio, interaction, model, representation, or compatibility question that ordinary delivery work does not fully answer.
- You can contribute domain expertise, code, data or assets with known rights, an audience or product context, research capacity, or funding.
- The work benefits from becoming reproducible, playable, measurable, publishable, transferable, or usable by musicians.
- You are prepared to define privacy, publication, IP, attribution, licensing, production support, and maintenance explicitly.
How partnerships work
Turn a promising question into a bounded project with credible evidence.
Frame
Name the question, musical stakes, prior work, contributions on both sides, smallest decisive result, and evidence the claim requires.
Build and study
Create the smallest useful combination of research software, model or data work, Engine integration, interface, benchmark, and musical evaluation.
Share or transfer
Publish findings, prepare a paper, sustain open infrastructure, transfer the result, productize it, license a relevant system, or stop cleanly.
Terms: Partner status does not automatically grant IP ownership, exclusivity, roadmap control, source or model access, publication rights, or a production-support promise. Funding, confidentiality, data use, attribution, licensing, dissemination, and maintenance are agreed per project.
Foundational research
The Engine helps research become playable software.
Its compiler, SDK, DSP, UI kit, project model, and qualification tools shorten the path from a finding to an instrument, interface, or DAW integration. Studio and the developer preview are two ways to experience that foundation—not the full scope of the lab.
Start a conversation
Bring the question worth answering together.
It can begin with a paper idea, a benchmark gap, a model or dataset, an existing product, a format or device, or a musical practice that current tools cannot yet express. We will help identify the smallest project that can produce a useful finding and a credible next step.