Riffmaw

In plain words: App: Riffmaw (music making)

Designed 3 min read 7 glossary terms 25 connected pages
The short version written for the Snayu team

Riffmaw is the planned music app: describe an idea, bring a reference or performance, hear alternatives and shape the music with AI help, through four roads: Forge, Perform, Jam and Produce.

Why read it: Read first before working on music generation, recording, jam sessions or mixing features.
  1. 1You choose what to keep; the assistant proposes material and edits within scope.
  2. 2Each road ends its own job; results pass forward as exact artifacts, never silently resumed.
  3. 3Owns music only: speech, dubbing, picture sound, visuals and publishing belong to other apps.

Designed Native app not built; only offline engineering probes exist (see Workbench).

Words on this page, decoded click one for its card

In this section 10 pages

Sources Riffmaw: reference material How Riffmaw admits specific audio files as references and marks an exact musical moment (for example a drop at 01:32) without turning the whole media library into influence. open → Sessions Riffmaw: recording and jamming How Riffmaw records performances (takes) and runs live jam sessions with human and machine players on shared, mapped clocks, preserving raw audio and MIDI before any editing. open → Music Hardware Boundary Riffmaw: audio hardware routes How Riffmaw admits an exact audio input route (interface, channel, clock, latency, monitoring) for a take, keeping the dry instrument signal separate from the processed sound you hear. open → Music Riffmaw: composing and mixing How Riffmaw develops material into an arrangement, mix and master: composing with models or humans, lyrics and sung language, live answer latency classes, autonomy levels, and mixing and mastering. open → Studio Riffmaw: production tools The production loop and the replaceable tool roles (planner, generator, analyzer, renderer, live host, mastering and so on), plus a dated map of candidate tools like MiniMax Music 3, DawDreamer and PipeWire. open → Workbench Riffmaw: build plan and probes The construction plan for an edit-and-listen music loop, plus a working offline engineering tool: render bounded recipes through DawDreamer, audition them, make guarded edits, and probe a local ComfyUI server. open → Voice boundary Riffmaw: vocal boundary Explains which vocals Riffmaw owns (singing, rap, musical spoken word, harmonies) and which belong elsewhere, and how performer consent and lyrics are bound to takes. open → Sound boundary Riffmaw: sound-effects boundary Explains that sound belongs to whichever app judges its intended role: musical textures and effects to Riffmaw, picture-bound effects to Broadcast, game sound to Foundry. open → Sync Riffmaw: music cue maps How Riffmaw publishes a MusicCueMap: exact musical events (drops, sections, entrances) on a declared timebase that visual or editing apps can respond to without editing the music. open → Package Riffmaw: sealed music bundle How Riffmaw seals accepted music, stems, mixes and masters into a MusicAssetBundle for other apps, and how it handles effects, recovery, revocation and deletion. open →
The original text Underlined words open their plain-English card.

Riffmaw is the music-making application being developed for LychD. Its intended experience is to describe an idea, bring a reference or performance, hear alternatives, and shape the music with AI assistance and approachable controls. The interface should offer detailed production controls when needed, while keeping small changes accessible without DAW expertise.

An assistant can help develop a brief, propose material or edits, and compare candidates within the requested scope. The MagusYou, the Operator chooses what to keep. A remembered drop, a guitar take, a sung line, or an empty session can all begin the craft; their contributions remain attributable.

Start with a musical idea

The intended interaction begins with a musical request, for example:

  • “Make a short psytrance sketch, with a quieter opening and a stronger bass entrance.”
  • “Turn this part down; let me compare it with the previous version.”
  • “Use this reference moment to explore a similar texture with different notes.”

Workbench develops that edit-and-listen experience. Its development operations describe the finite engineering tools available to exercise today; they do not yet supply the native application or an AI listening assistant. State of Work keeps that implementation boundary.

Ways of making

Road Pattern What can finish here
Forge riffmaw.forge_track@2 Offline composition/generation from a brief, references, lyrics, or empty session: MusicCandidateSet@1 or MusicFindingSet@1.
Perform riffmaw.capture_take@2 Explicitly armed audio/MIDI from played, sung, rapped, musical-spoken-word, or programmed performance: PerformanceTake@1 and observations/findings.
Jam riffmaw.open_jam@2 Performers and admitted machine players answering against mapped clocks: recoverable MusicalPerformanceSession@1 or findings.
Produce riffmaw.forge_music_bundle@2 Bounded arrangement, mix, master, probes and listening review: MusicAssetBundle@1, candidates, findings, or refusal.

Each road ends by settling its own InvocationJob Request. Carry exact artifacts into later forward work without nesting or silently resuming another PatternWorkflow. A future ScrollWorkflow Version may combine separately named capture/generation/render/review placements, but must do so under its own revision, Resolution Lock, and authority ceiling.

The passage between roads preserves their separate decisions: file access does not arm an input, a take does not admit training, a candidate mix is not master approval, and export does not publish.

The work

Find the material in Sources, down to the exact musical moment. When people or machines perform, Sessions preserves their takes and shared time; Hardware admits the capture routes they use. Music develops that material into an arrangement and master candidate.

Choose machinery through Studio, which keeps tool roles and dated candidates separate from the craft. For uncertain roles, Voice distinguishes musical performance from speech, and Sound follows what belongs to music. Sync offers cues for another work; Package seals the musical handoff.

Reference contract

riffmaw.music revision 2 accepts a music brief plus explicitly admitted references, lyrics, takes, MIDI, scores, samples, or live inputs. Alongside the road-specific returns, it can issue MusicCueMap@1. It stops before ordinary dialogue, spoken replacement/localization, picture sound, visuals, final audiovisual assembly, impersonation, and publication. Language Edition, Broadcast, Atelier, and their technical providers retain those offices.

Prompt, lyric, requested sung language, and interface locale remain separate. Exact model profiles must prove their musical/language behavior; a local ComfyUI or other SoulstoneLocal Model binding and a PortalCloud Model API are independently evidenced implementations. Music owns the request; Studio records candidates.

Revision continuity

Earlier designs and their exact successors

riffmaw.audio revision 1, its Sonic* family, and former riffmaw.forge_track@1, riffmaw.capture_take@1, riffmaw.open_jam@1, riffmaw.index_sources@1, and riffmaw.mark_moment@1 are retired without reinterpretation. New identity riffmaw.music revision 1 used the five PatternWorkflow names at @2. Revision 2 advances principal riffmaw.forge_music_bundle to @2 for model-native pre-execution music authority and semantic admission. MusicAssetBundle@1 already accommodates those exact provenance receipts; other music PatternsWorkflows remain unchanged. No registry, RunJob Record, or stored application record used the retired designs, so no executable migration exists.

Audio · Workflow · Composition Portfolio