@linto-ai/transcript-ui-plugin-transcription-editor

Part of @linto-ai/transcript-ui — see there for the full package list and quick start.

Collaborative editing of turns and speakers: rename, split, merge, and reassign, with per-turn locks and server-authoritative sync. This plugin never computes the retimed truth itself — every option below is optimistic-local-then-server-broadcast: the edit applies instantly in the UI, your callback pushes it to the backend, and the backend's broadcast (applied back through core.transcriptionEditor.apply*) is what actually sticks.

Usage

import { createTranscriptionEditorPlugin } from "@linto-ai/transcript-ui-plugin-transcription-editor"

core.use(
  createTranscriptionEditorPlugin({
    saveTurn: async ({ translationId, turnId, text }) => {
      const ok = await api.saveTurn(translationId, turnId, text)
      return { ok }
    },
    lockTurn: async ({ translationId, turnId }) => {
      const holder = await api.lockTurn(translationId, turnId)
      return { ok: !holder, holder }
    },
    unlockTurn: async ({ translationId, turnId }) => {
      await api.unlockTurn(translationId, turnId)
      return { ok: true }
    },
    // splitTurn, deleteTurn, mergeTurns, updateTurnSpeaker, renameSpeaker,
    // replaceSpeaker, refetchTranslation — same shape, all optional.
  }),
)

Apply what the backend broadcasts to other clients (e.g. over a websocket) the same way:

socket.on("editor:turn_updated", (update) => {
  core.transcriptionEditor?.applyTurnUpdate(update)
})

Locks broadcast by the server are pushed with setLocks/setTurnLock/clearTurnLock; a reconnect re-ack that reports a version gap is handled by calling your refetchTranslation and then reconcileVersions.

In production

ConversationsTranscription.vue (LinTO Studio) wires every callback above to its WebSocket API.