app_dubbing
ChatGPTCreate a voiceover from supplied text and a real reference_audio_id. Poll the returned session_id with voxstudio.dub.result.
app_knowledge
ChatGPTBegin an evidence question or show a previous workspace. begin requires a nonempty query and fresh UUID request_id; preserve the ID on retry.
app_transcription
ChatGPTSubmit an HTTPS attachment with media_file; prepare a local upload with file_name and mime_type; or complete an uploaded session with session_id. Follow next_call with the exact session_id.
knowledge_complete_turn
ChatGPTFreeze the answer sources for this question after reads complete. Only this turn's successful reads and non-search observations may be cited. Does not generate an answer.
knowledge_preview
ChatGPTRead an owned media session's audio through a bounded MCP resource. source_id is a knowledge ID; the returned session_id is the real media ID.
knowledge_update_view
ChatGPTSave reading focus, pin or next-question scope without modifying source documents. Use the current view revision and preserve request_id on retry.
knowledge_workspace_state
ChatGPTRead current presentation state or an immutable historical question snapshot.
voxstudio_dub_reference_list
ChatGPTList the current user's accessible private and public dubbing reference audios for a read language.
voxstudio_dub_result
ChatGPTReturn dubbing job status and a signed audio URL when the dub audio is ready.
voxstudio_knowledge_aggregate
ChatGPTCount and group the full authorized title-filtered inventory; report unknown media durations separately.
voxstudio_knowledge_fetch
ChatGPTRead current canonical text, summary, existing derived navigation data, metadata or an evidence locator. For short text read continuously. For long text use search/find_text or targeted start/end ranges for the requested dimensions; follow next_cursor when the task requires exhaustive coverage. A partial prefix cannot establish absence elsewhere in the material.
voxstudio_knowledge_find_text
ChatGPTLiteral casefold/whitespace matching mapped to original Unicode codepoint spans, with occurrence count and pagination.
voxstudio_knowledge_list_sources
ChatGPTPage through the entire authorized source directory. Query is a literal title filter.
voxstudio_knowledge_methods
ChatGPTRead the published method catalog or one method. Methods guide research without replacing the core tools.
voxstudio_knowledge_search
ChatGPTDiscover sources semantically using canonical text, summaries and derived navigation data, or retrieve bounded canonical passage candidates. Use list_sources for literal titles. Candidates are not a complete inventory.
voxstudio_media_preview_url
ChatGPTGet a short-lived signed audio or video preview URL.
voxstudio_session_export
ChatGPTExport completed ASR transcript/subtitles or TTS audio as a signed download URL.
voxstudio_session_get
ChatGPTGet one owned media session.
voxstudio_session_list_quotes
ChatGPTList the user's saved quotes for one owned session. Each quote is a frozen snapshot with speakers, time precision, human verification status and whether the transcript changed since it was saved. Read-only: quotes can only be created, edited or verified by the user in VoxStudio.
voxstudio_transcript_context
ChatGPTGet neighboring transcript context around a time range.
voxstudio_workspace
ChatGPTOpen the session and evidence workspace. One workspace per chat; no answer is generated.