Integrate Maru into your product
For DAWs, online music services and existing plug-in makers: what Maru can add to your product, how each part works, and how to start with one focused integration.
Maru is also a foundation other products can build on. A browser DAW can offer playable instruments and effects. A music-learning or sample service can add sound generation without running models. A plug-in company can give its interfaces, presets and release process a modern, agent-friendly workflow. This page explains what you can integrate, how each part works and how to start.
Who this is for
| You run | What Maru can add |
|---|---|
| A DAW or music app | Playable instruments, effects and MIDI tools embedded in your app; native plug-in builds of the same modules for desktop DAWs |
| An online service (learning, samples, collaboration, creative AI) | Hosted sound generation, stems and MIDI extraction for your users; embeddable instruments and a sampler; musical theory computation |
| An existing plug-in line | One project for sound, interface and presets; interfaces ported a control at a time; agent-assisted editing; licensing and delivery |
| An AI or agent product | Maru's operations as MCP tools, so your users' assistants can create projects, generate sounds and read guides |
| A controller maker | Instruments that use every dimension of your controller (MPE, grids, gestures), ready for your owners |
| A hardware maker | Instruments and effects authored once and exported to small devices (early access) |
| A sample or content publisher | Libraries that become playable, sellable instruments with licensing and clear provenance |
| A model or research lab | A playable front end for your model, with nothing for end users to host |
| An education or creator platform | Theory, practice and creation tools that your users' agents can drive |
Why build on it
- The library is already there. Sixty named effects, a library of instruments and MIDI tools, a sampler and interface controls are available as editable source. See What you can build.
- One source for every destination. The same module plays in a browser and builds as a native plug-in, so you maintain one sound and one interface instead of several ports. See How it works.
- No model hosting. Generation runs on Maru's servers. With a registered app, your users generate on their own accounts and credits, after approving your app on Maru's consent page. See Generate audio.
- Agents work through real operations. Every operation is a CLI command, an MCP tool and a typed call with one schema and one permission model, so agent work stays inspectable.
- Licensing and checkout included. Sell through Maru, or keep your own distribution. See Sell and share.
How each part works for a partner
Embedding. A module host loads a module's release into your web app, connects its audio and MIDI, and shows its interface. Your app keeps its own transport, files and accounts; the module owns its sound and parameters. Start with one instrument or effect, then connect several in a rack.
Interfaces. Interfaces are browser code bound to parameters, so they can be restyled to match your product, or an existing panel can be ported one control at a time while the sound stays the same. See Module UI.
Generation. Your server can call the generation API with a token, or your app can sign people in with a registered creator app so each person generates on their own account. Results are ordinary audio your product loads through its existing sample workflow.
Agents. The hosted connector and the hosted operations API expose projects, builds, guides and generation to any MCP client or script, with access the account owner approves. See Connect your AI assistant.
Plug-ins and delivery. Modules build as native plug-ins for the formats they support, with optional license activation. Formats are qualified per module and platform, and each release lists what it includes.
Start with one focused integration
We work with partners on one bounded project first, agreed around a real musical result: for example embedding one instrument in your app, adding a Generate action to an existing tool, porting one plug-in interface or automating one release step. Together we agree the user, the intended host and how we will judge success before expanding.
Maru is in beta, and some paths, such as the local toolchain and native release builds, are available to selected partners and developers. Tell us what you have in mind on Work with us, or read the Engine overview for the architecture.
Technical details
| Integration | Entry point | Guide |
|---|---|---|
| Embed a module in React | Module player and face host | Host a module in React |
| Generation from your server | createMaruClient from maruaudio/api | Add generative audio to your app |
| Generation for signed-in users | A registered creator app with PKCE sign-in | Add generative audio to your app |
| Agent and script access | Hosted connector at https://maruaudio.com/mcp; POST /api/v1/ops/<name> | Use Maru from your agent, MCP tools |
| Music theory | The Symbolic library | Symbolic library |
| Interfaces | UI kit controls and the face contract | Module UI |
| Native plug-ins and licensing | Release builds and plug-in licensing | Native plugins, Plug-in licensing |
| Comparing approaches | Maru alongside JUCE, Cmajor and RNBO | Choosing Maru, comparison |
The SDK, CLI and generated plug-in shells are not published under a general open-source license; see Licensing for current terms.