{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://raw.githubusercontent.com/shruggietech/insonic/v1.0.0/schemas/v1.0.0/import-manifest.schema.json",
  "title": "insonic import manifest",
  "description": "Batch import input for the CLI and its GUI wrapper. Syntax validation precedes application validation of source access, calendar values, timezone/DST policy and admission. No approval per item is required for already configured routing.",
  "type": "object",
  "properties": {
    "schema_version": {
      "const": "1.0.0",
      "description": "Version of the insonic document contract."
    },
    "kind": {
      "const": "import-manifest",
      "description": "Document discriminator used by the master registry."
    },
    "defaults": {
      "type": "object",
      "description": "Recording-date defaults and policies. Exact timestamp, calendar date and approximate UTC range are mutually exclusive.",
      "properties": {
        "originated_at": {
          "$ref": "common.schema.json#/$defs/wallTimestamp",
          "description": "Entered recording timestamp, distinct from import, publication and file modification times."
        },
        "originated_on": {
          "$ref": "common.schema.json#/$defs/calendarDate",
          "description": "Entered recording date when the recording time is unknown."
        },
        "timezone": {
          "$ref": "common.schema.json#/$defs/timezone",
          "description": "Timezone for unzoned recording timestamps; local is resolved once for the batch."
        },
        "dst_fold": {
          "type": "string",
          "enum": [
            "earlier",
            "later"
          ],
          "description": "Explicit choice of offset for an ambiguous repeated daylight-saving wall time."
        },
        "dst_gap": {
          "type": "string",
          "const": "shift-forward",
          "description": "Explicit policy to resolve a nonexistent wall time by shifting forward; absent means no silent correction."
        },
        "copy": {
          "type": "boolean",
          "default": true,
          "description": "True admits a managed copy into the selected artifact store; false keeps an explicitly tracked local reference."
        },
        "preset": {
          "type": "string",
          "minLength": 1,
          "description": "Configured processing/import preset name or ID; does not embed provider configuration.",
          "examples": [
            "local"
          ]
        },
        "extensions": {
          "$ref": "common.schema.json#/$defs/extensions",
          "description": "Optional nonsecret input annotations."
        },
        "date_precedence": {
          "type": "string",
          "enum": [
            "owner-first",
            "embedded-first",
            "filesystem-fallback"
          ],
          "description": "Versioned recording-date precedence rule. Filesystem modification becomes a recording-date candidate only under the explicit filesystem-fallback rule."
        },
        "originated_earliest": {
          "$ref": "common.schema.json#/$defs/utcInstant",
          "description": "Inclusive earliest recording instant for an owner-supplied approximate date range. Preserve the exact UTC timestamp and integer clock."
        },
        "originated_latest": {
          "$ref": "common.schema.json#/$defs/utcInstant",
          "description": "Inclusive latest recording instant for an owner-supplied approximate date range. Application validation checks bound ordering."
        },
        "local_http": {
          "type": "boolean",
          "default": false,
          "description": "Explicitly elect HTTP credential transport for a loopback endpoint only. Credentials otherwise require HTTPS; unauthenticated HTTP sources remain supported."
        },
        "acquisition_max_bytes": {
          "type": "integer",
          "minimum": 1,
          "default": 68719476736,
          "description": "Maximum acquired original bytes, default 64 GiB. Declared and streamed lengths are bounded before publication."
        },
        "acquisition_timeout_ms": {
          "type": "integer",
          "minimum": 1,
          "maximum": 9223372036854,
          "default": 600000,
          "description": "Positive total acquisition deadline in milliseconds, default ten minutes. Per-item values override batch defaults."
        },
        "transcript_max_bytes": {
          "type": "integer",
          "minimum": 1,
          "maximum": 16777216,
          "default": 16777216,
          "description": "Independent transcript acquisition and document byte ceiling (maximum 16 MiB)."
        },
        "transcript_timeout_ms": {
          "type": "integer",
          "minimum": 1,
          "maximum": 600000,
          "default": 120000,
          "description": "Independent transcript acquisition timeout in milliseconds."
        },
        "attribution": {
          "type": "string",
          "enum": [
            "auto",
            "native",
            "off",
            "diarize"
          ],
          "default": "auto",
          "description": "Preserve existing assignments; derive structured participation only when none exist. Diarize explicitly elects configured acoustic processing without recognition."
        },
        "transcript_format": {
          "type": "string",
          "enum": [
            "cueson",
            "srt",
            "vtt",
            "ass",
            "ssa"
          ],
          "description": "Explicit native or Cue JSON format hint validated against input bytes."
        },
        "replace_transcript": {
          "type": "boolean",
          "default": false,
          "description": "Elect replacement when the target already has a current transcript."
        },
        "subtitle_stream_index": {
          "type": "integer",
          "minimum": 0,
          "description": "Absolute embedded subtitle stream index to select."
        },
        "subtitle_language": {
          "type": "string",
          "minLength": 1,
          "description": "Case-insensitive embedded subtitle language filter."
        },
        "diarization_model_id": {
          "$ref": "common.schema.json#/$defs/modelReference",
          "description": "Shared model reference; submission freezes an exact compatible target before acquisition and processing."
        },
        "replace_audio": {
          "type": "boolean",
          "description": "Explicit election to replace canonical audio at an existing stable recording."
        },
        "existing_transcript": {
          "enum": [
            "keep",
            "clear",
            "replace"
          ],
          "description": "Required document election when replacement has a current transcript."
        },
        "transcript_applies": {
          "type": "boolean",
          "description": "Affirm that the current transcript applies to elected replacement audio."
        },
        "existing_roster": {
          "enum": [
            "retain",
            "clear"
          ],
          "description": "Explicit declared-roster election during audio replacement."
        },
        "known_speakers": {
          "type": "array",
          "maxItems": 1000,
          "items": {
            "type": "string",
            "minLength": 1,
            "maxLength": 512
          },
          "description": "Initial declared identities by UUID or unique exact canonical name/active alias. Empty declares an empty roster."
        }
      },
      "not": {
        "type": "object",
        "anyOf": [
          {
            "required": [
              "originated_at",
              "originated_on"
            ]
          },
          {
            "required": [
              "originated_at",
              "originated_earliest"
            ]
          },
          {
            "required": [
              "originated_at",
              "originated_latest"
            ]
          },
          {
            "required": [
              "originated_on",
              "originated_earliest"
            ]
          },
          {
            "required": [
              "originated_on",
              "originated_latest"
            ]
          }
        ]
      },
      "additionalProperties": false
    },
    "items": {
      "type": "array",
      "minItems": 1,
      "description": "Inputs in deterministic admission order, with independent partial-result reporting.",
      "items": {
        "type": "object",
        "description": "Recording-date defaults and policies. Exact timestamp, calendar date and approximate UTC range are mutually exclusive.",
        "properties": {
          "source": {
            "type": "string",
            "minLength": 1,
            "description": "Local source path or acquisition-adapter source locator. Credentials and temporary signed URLs must be resolved separately.",
            "examples": [
              "audio/session.wav"
            ]
          },
          "subtitle": {
            "type": "string",
            "minLength": 1,
            "description": "Compatibility alias for transcript, with the same independent acquisition and validation contract.",
            "examples": [
              "video/session.srt"
            ]
          },
          "originated_at": {
            "$ref": "common.schema.json#/$defs/wallTimestamp",
            "description": "Entered recording timestamp, distinct from import, publication and file modification times."
          },
          "originated_on": {
            "$ref": "common.schema.json#/$defs/calendarDate",
            "description": "Entered recording date when the recording time is unknown."
          },
          "timezone": {
            "$ref": "common.schema.json#/$defs/timezone",
            "description": "Timezone for unzoned recording timestamps; local is resolved once for the batch."
          },
          "dst_fold": {
            "type": "string",
            "enum": [
              "earlier",
              "later"
            ],
            "description": "Explicit choice of offset for an ambiguous repeated daylight-saving wall time."
          },
          "dst_gap": {
            "type": "string",
            "const": "shift-forward",
            "description": "Explicit policy to resolve a nonexistent wall time by shifting forward; absent means no silent correction."
          },
          "copy": {
            "type": "boolean",
            "default": true,
            "description": "True admits a managed copy into the selected artifact store; false keeps an explicitly tracked local reference."
          },
          "preset": {
            "type": "string",
            "minLength": 1,
            "description": "Configured processing/import preset name or ID; does not embed provider configuration.",
            "examples": [
              "local"
            ]
          },
          "extensions": {
            "$ref": "common.schema.json#/$defs/extensions",
            "description": "Optional nonsecret input annotations."
          },
          "title": {
            "type": "string",
            "minLength": 1,
            "description": "Optional library display title; source identity still comes from verified bytes."
          },
          "new_entry": {
            "type": "boolean",
            "default": false,
            "description": "Explicitly create another library entry even when this source was already admitted. Identical bytes can share the source asset."
          },
          "credential_id": {
            "$ref": "common.schema.json#/$defs/uuid",
            "description": "Workspace-scoped acquisition credential reference, resolved privately by the configured adapter."
          },
          "acquisition_adapter": {
            "type": "string",
            "minLength": 1,
            "description": "Explicit configured acquisition adapter; absent selects the source URI scheme without changing providers on failure."
          },
          "date_precedence": {
            "type": "string",
            "enum": [
              "owner-first",
              "embedded-first",
              "filesystem-fallback"
            ],
            "description": "Versioned recording-date precedence rule. Filesystem modification becomes a recording-date candidate only under the explicit filesystem-fallback rule."
          },
          "originated_earliest": {
            "$ref": "common.schema.json#/$defs/utcInstant",
            "description": "Inclusive earliest recording instant for an owner-supplied approximate date range. Preserve the exact UTC timestamp and integer clock."
          },
          "originated_latest": {
            "$ref": "common.schema.json#/$defs/utcInstant",
            "description": "Inclusive latest recording instant for an owner-supplied approximate date range. Application validation checks bound ordering."
          },
          "local_http": {
            "type": "boolean",
            "default": false,
            "description": "Explicitly elect HTTP credential transport for a loopback endpoint only. Credentials otherwise require HTTPS; unauthenticated HTTP sources remain supported."
          },
          "acquisition_max_bytes": {
            "type": "integer",
            "minimum": 1,
            "default": 68719476736,
            "description": "Maximum acquired original bytes, default 64 GiB. Declared and streamed lengths are bounded before publication."
          },
          "acquisition_timeout_ms": {
            "type": "integer",
            "minimum": 1,
            "maximum": 9223372036854,
            "default": 600000,
            "description": "Positive total acquisition deadline in milliseconds, default ten minutes. Per-item values override batch defaults."
          },
          "transcript_max_bytes": {
            "type": "integer",
            "minimum": 1,
            "maximum": 16777216,
            "default": 16777216,
            "description": "Independent transcript acquisition and document byte ceiling (maximum 16 MiB)."
          },
          "transcript_timeout_ms": {
            "type": "integer",
            "minimum": 1,
            "maximum": 600000,
            "default": 120000,
            "description": "Independent transcript acquisition timeout in milliseconds."
          },
          "attribution": {
            "type": "string",
            "enum": [
              "auto",
              "native",
              "off",
              "diarize"
            ],
            "default": "auto",
            "description": "Preserve existing assignments; derive structured participation only when none exist. Diarize explicitly elects configured acoustic processing without recognition."
          },
          "transcript_format": {
            "type": "string",
            "enum": [
              "cueson",
              "srt",
              "vtt",
              "ass",
              "ssa"
            ],
            "description": "Explicit native or Cue JSON format hint validated against input bytes."
          },
          "replace_transcript": {
            "type": "boolean",
            "default": false,
            "description": "Elect replacement when the target already has a current transcript."
          },
          "subtitle_stream_index": {
            "type": "integer",
            "minimum": 0,
            "description": "Absolute embedded subtitle stream index to select."
          },
          "subtitle_language": {
            "type": "string",
            "minLength": 1,
            "description": "Case-insensitive embedded subtitle language filter."
          },
          "diarization_model_id": {
            "$ref": "common.schema.json#/$defs/modelReference",
            "description": "Shared model reference; submission freezes an exact compatible target before acquisition and processing."
          },
          "transcript": {
            "type": "string",
            "minLength": 1,
            "description": "Independent local or authorized acquisition locator for native subtitles or Cue JSON. Explicit invalid input fails; it never falls back to embedded/generated text."
          },
          "record": {
            "type": "string",
            "minLength": 1,
            "description": "Existing recording UUID or unambiguous exact title for standalone transcript admission."
          },
          "transcript_credential_id": {
            "$ref": "common.schema.json#/$defs/uuid",
            "description": "Independent credential UUID for transcript acquisition."
          },
          "transcript_adapter": {
            "type": "string",
            "minLength": 1,
            "description": "Explicit independent transcript acquisition adapter."
          },
          "expected_revision": {
            "type": "integer",
            "minimum": 1,
            "description": "Optional observed target media revision for standalone transcript admission."
          },
          "expected_recording_revision": {
            "type": "integer",
            "minimum": 0,
            "description": "Optional observed current document revision (zero means absent)."
          },
          "replace_audio": {
            "type": "boolean",
            "description": "Explicit election to replace canonical audio at an existing stable recording."
          },
          "existing_transcript": {
            "enum": [
              "keep",
              "clear",
              "replace"
            ],
            "description": "Required document election when replacement has a current transcript."
          },
          "transcript_applies": {
            "type": "boolean",
            "description": "Affirm that the current transcript applies to elected replacement audio."
          },
          "existing_roster": {
            "enum": [
              "retain",
              "clear"
            ],
            "description": "Explicit declared-roster election during audio replacement."
          },
          "known_speakers": {
            "type": "array",
            "maxItems": 1000,
            "items": {
              "type": "string",
              "minLength": 1,
              "maxLength": 512
            },
            "description": "Initial declared identities by UUID or unique exact canonical name/active alias. Empty declares an empty roster."
          },
          "kind": {
            "enum": [
              "media",
              "transcript"
            ],
            "description": "Media with record targets audio; transcript preserves the standalone source alias."
          },
          "expected_roster_revision": {
            "type": "integer",
            "minimum": 0,
            "description": "Observed roster revision, zero for absent."
          }
        },
        "required": [],
        "not": {
          "type": "object",
          "anyOf": [
            {
              "required": [
                "originated_at",
                "originated_on"
              ]
            },
            {
              "required": [
                "originated_at",
                "originated_earliest"
              ]
            },
            {
              "required": [
                "originated_at",
                "originated_latest"
              ]
            },
            {
              "required": [
                "originated_on",
                "originated_earliest"
              ]
            },
            {
              "required": [
                "originated_on",
                "originated_latest"
              ]
            },
            {
              "required": [
                "transcript",
                "subtitle"
              ]
            }
          ]
        },
        "additionalProperties": false,
        "oneOf": [
          {
            "required": [
              "source",
              "kind"
            ],
            "properties": {
              "kind": {
                "const": "media",
                "description": "Selected input discriminator."
              }
            },
            "not": {
              "required": [
                "transcript",
                "subtitle"
              ]
            }
          },
          {
            "required": [
              "source"
            ],
            "not": {
              "anyOf": [
                {
                  "required": [
                    "record"
                  ]
                },
                {
                  "required": [
                    "kind"
                  ],
                  "properties": {
                    "kind": {
                      "const": "media",
                      "description": "Selected input discriminator."
                    }
                  }
                }
              ]
            }
          },
          {
            "required": [
              "record"
            ],
            "not": {
              "required": [
                "kind"
              ],
              "properties": {
                "kind": {
                  "const": "media",
                  "description": "Selected input discriminator."
                }
              }
            },
            "oneOf": [
              {
                "required": [
                  "source"
                ],
                "not": {
                  "anyOf": [
                    {
                      "required": [
                        "transcript"
                      ]
                    },
                    {
                      "required": [
                        "subtitle"
                      ]
                    }
                  ]
                }
              },
              {
                "required": [
                  "transcript"
                ],
                "not": {
                  "anyOf": [
                    {
                      "required": [
                        "source"
                      ]
                    },
                    {
                      "required": [
                        "subtitle"
                      ]
                    }
                  ]
                }
              },
              {
                "required": [
                  "subtitle"
                ],
                "not": {
                  "anyOf": [
                    {
                      "required": [
                        "source"
                      ]
                    },
                    {
                      "required": [
                        "transcript"
                      ]
                    }
                  ]
                }
              }
            ]
          }
        ],
        "dependentRequired": {
          "expected_revision": [
            "record"
          ],
          "expected_recording_revision": [
            "record"
          ],
          "expected_roster_revision": [
            "record"
          ]
        }
      },
      "maxItems": 10000
    },
    "extensions": {
      "$ref": "common.schema.json#/$defs/extensions",
      "description": "Optional namespaced additions preserved without changing the core contract."
    }
  },
  "required": [
    "schema_version",
    "kind",
    "items"
  ],
  "additionalProperties": false,
  "examples": [
    {
      "schema_version": "1.0.0",
      "kind": "import-manifest",
      "defaults": {
        "copy": true,
        "timezone": "America/New_York",
        "preset": "local"
      },
      "items": [
        {
          "source": "audio/session.wav",
          "originated_at": "2026-10-04T14:30:00"
        },
        {
          "source": "video/session.mp4",
          "originated_on": "2026-10-03",
          "subtitle": "video/session.srt"
        }
      ]
    },
    {
      "kind": "import-manifest",
      "schema_version": "1.0.0",
      "items": [
        {
          "source": "audio/new.flac",
          "known_speakers": []
        },
        {
          "kind": "media",
          "source": "audio/replacement.flac",
          "record": "44444444-4444-4444-8444-444444444444",
          "replace_audio": true,
          "existing_transcript": "keep",
          "transcript_applies": true,
          "existing_roster": "retain"
        }
      ]
    }
  ],
  "$defs": {
    "options": {
      "type": "object",
      "description": "Recording-date defaults and policies. Exact timestamp, calendar date and approximate UTC range are mutually exclusive.",
      "properties": {
        "originated_at": {
          "$ref": "common.schema.json#/$defs/wallTimestamp",
          "description": "Entered recording timestamp, distinct from import, publication and file modification times."
        },
        "originated_on": {
          "$ref": "common.schema.json#/$defs/calendarDate",
          "description": "Entered recording date when the recording time is unknown."
        },
        "timezone": {
          "$ref": "common.schema.json#/$defs/timezone",
          "description": "Timezone for unzoned recording timestamps; local is resolved once for the batch."
        },
        "dst_fold": {
          "type": "string",
          "enum": [
            "earlier",
            "later"
          ],
          "description": "Explicit choice of offset for an ambiguous repeated daylight-saving wall time."
        },
        "dst_gap": {
          "type": "string",
          "const": "shift-forward",
          "description": "Explicit policy to resolve a nonexistent wall time by shifting forward; absent means no silent correction."
        },
        "copy": {
          "type": "boolean",
          "default": true,
          "description": "True admits a managed copy into the selected artifact store; false keeps an explicitly tracked local reference."
        },
        "preset": {
          "type": "string",
          "minLength": 1,
          "description": "Configured processing/import preset name or ID; does not embed provider configuration.",
          "examples": [
            "local"
          ]
        },
        "extensions": {
          "$ref": "common.schema.json#/$defs/extensions",
          "description": "Optional nonsecret input annotations."
        },
        "date_precedence": {
          "type": "string",
          "enum": [
            "owner-first",
            "embedded-first",
            "filesystem-fallback"
          ],
          "description": "Versioned recording-date precedence rule. Filesystem modification becomes a recording-date candidate only under the explicit filesystem-fallback rule."
        },
        "originated_earliest": {
          "$ref": "common.schema.json#/$defs/utcInstant",
          "description": "Inclusive earliest recording instant for an owner-supplied approximate date range. Preserve the exact UTC timestamp and integer clock."
        },
        "originated_latest": {
          "$ref": "common.schema.json#/$defs/utcInstant",
          "description": "Inclusive latest recording instant for an owner-supplied approximate date range. Application validation checks bound ordering."
        },
        "local_http": {
          "type": "boolean",
          "default": false,
          "description": "Explicitly elect HTTP credential transport for a loopback endpoint only. Credentials otherwise require HTTPS; unauthenticated HTTP sources remain supported."
        },
        "acquisition_max_bytes": {
          "type": "integer",
          "minimum": 1,
          "default": 68719476736,
          "description": "Maximum acquired original bytes, default 64 GiB. Declared and streamed lengths are bounded before publication."
        },
        "acquisition_timeout_ms": {
          "type": "integer",
          "minimum": 1,
          "maximum": 9223372036854,
          "default": 600000,
          "description": "Positive total acquisition deadline in milliseconds, default ten minutes. Per-item values override batch defaults."
        },
        "transcript_max_bytes": {
          "type": "integer",
          "minimum": 1,
          "maximum": 16777216,
          "default": 16777216,
          "description": "Independent transcript acquisition and document byte ceiling (maximum 16 MiB)."
        },
        "transcript_timeout_ms": {
          "type": "integer",
          "minimum": 1,
          "maximum": 600000,
          "default": 120000,
          "description": "Independent transcript acquisition timeout in milliseconds."
        },
        "attribution": {
          "type": "string",
          "enum": [
            "auto",
            "native",
            "off",
            "diarize"
          ],
          "default": "auto",
          "description": "Preserve existing assignments; derive structured participation only when none exist. Diarize explicitly elects configured acoustic processing without recognition."
        },
        "transcript_format": {
          "type": "string",
          "enum": [
            "cueson",
            "srt",
            "vtt",
            "ass",
            "ssa"
          ],
          "description": "Explicit native or Cue JSON format hint validated against input bytes."
        },
        "replace_transcript": {
          "type": "boolean",
          "default": false,
          "description": "Elect replacement when the target already has a current transcript."
        },
        "subtitle_stream_index": {
          "type": "integer",
          "minimum": 0,
          "description": "Absolute embedded subtitle stream index to select."
        },
        "subtitle_language": {
          "type": "string",
          "minLength": 1,
          "description": "Case-insensitive embedded subtitle language filter."
        },
        "diarization_model_id": {
          "$ref": "common.schema.json#/$defs/modelReference",
          "description": "Shared model reference; submission freezes an exact compatible target before acquisition and processing."
        },
        "replace_audio": {
          "type": "boolean",
          "description": "Explicit election to replace canonical audio at an existing stable recording."
        },
        "existing_transcript": {
          "enum": [
            "keep",
            "clear",
            "replace"
          ],
          "description": "Required document election when replacement has a current transcript."
        },
        "transcript_applies": {
          "type": "boolean",
          "description": "Affirm that the current transcript applies to elected replacement audio."
        },
        "existing_roster": {
          "enum": [
            "retain",
            "clear"
          ],
          "description": "Explicit declared-roster election during audio replacement."
        },
        "known_speakers": {
          "type": "array",
          "maxItems": 1000,
          "items": {
            "type": "string",
            "minLength": 1,
            "maxLength": 512
          },
          "description": "Initial declared identities by UUID or unique exact canonical name/active alias. Empty declares an empty roster."
        }
      },
      "not": {
        "type": "object",
        "anyOf": [
          {
            "required": [
              "originated_at",
              "originated_on"
            ]
          },
          {
            "required": [
              "originated_at",
              "originated_earliest"
            ]
          },
          {
            "required": [
              "originated_at",
              "originated_latest"
            ]
          },
          {
            "required": [
              "originated_on",
              "originated_earliest"
            ]
          },
          {
            "required": [
              "originated_on",
              "originated_latest"
            ]
          }
        ]
      },
      "additionalProperties": false
    }
  }
}