- Remove stage-tamagotchi, stage-pocket, server, and engine workspaces with their workflows, addons, and docs
- Switch the toolchain from pnpm to bun, replace lockfiles with bun.lock
- Rewrite remaining product references to Moeka

2026-09-14 16:40:49 +07:00

src

fix(ci): remove incomplete release tests (#2407 )

2026-08-30 03:11:51 +08:00

package.json

chore: prune unused apps and services for the moeka fork

2026-09-14 16:40:49 +07:00

README.md

chore: prune unused apps and services for the moeka fork

2026-09-14 16:40:49 +07:00

tsconfig.json

fix(stage-tamagotchi): bad import & lint

2025-10-20 00:47:34 +08:00

tsdown.config.ts

Revert "style: lint"

2026-08-26 20:13:10 +08:00

vitest.config.ts

fix(pipelines-audio): no longer concurrently tts

2026-03-31 15:51:24 +08:00

README.md

@proj-airi/pipelines-audio

Shared audio-pipeline orchestration for Moeka. The package owns reusable streaming, playback, text-chunking, and transcript-buffering policies without depending on an application UI.

Use it for

Building and scheduling speech playback pipelines.
Parsing streaming-control events.
Grouping nearby ASR fragments with createTranscriptBuffer before a product sends one spoken turn downstream.

Do not use it for

Vue or Electron lifecycle state.
Provider credentials and product-specific error UI.
Raw audio encoding utilities, which belong in @proj-airi/audio.

Transcript buffering

import { createTranscriptBuffer } from '@proj-airi/pipelines-audio'

const buffer = createTranscriptBuffer({
  flushDelayMs: 1200,
  flush: async text => sendToChat(text),
})

buffer.push('hello')
buffer.push('world')
await buffer.dispose()