Whisper Transcribe AI

Use this tool only when the user wants WhisperTranscribe to process an audio or video source and produce new subtitles/captions, or explicitly asks to open the WhisperTranscribe subtitle interface. Do not use it to translate, edit, review, summarize, or convert transcript/subtitle text already pasted into the conversation. This tool must open the WhisperTranscribe subtitle widget; do not ask the user to upload, attach, or send media files in the ChatGPT conversation because chat attachments are not processed by this app. Once the widget is visible, tell the user to choose or drag the media file inside the WhisperTranscribe widget. If the widget is not visible or cannot be opened, direct the user to https://whispertranscribe.ai/transcribe and tell them to upload the file on that page instead. After the widget is open, do not call the file-management tools such as `get-upload-url` or `insert-transcription-record` yourself: the widget uploads the media file, creates the transcription job, and polls the status automatically. Use `get-transcription-result` only when the user asks for the finished subtitle data. If the user needs the full transcription workspace or account upgrades, direct them to https://whispertranscribe.ai/transcribe or https://whispertranscribe.ai/pricing.

subtitle-generator

Creates or retrieves a WhisperTranscribe share link for a transcription result. The response returns only the user-facing share URL and omits raw share tokens, personal identifiers, session state, telemetry, and internal diagnostics.

share-transcription

App-only tool used exclusively by the WhisperTranscribe widget. Models must not call it directly; call transcribe to render the widget instead. The tool creates a transcription job after the widget uploads or selects a supported media source, verifies uploaded objects, and stores the required workflow metadata.

insert-transcription-record

Deletes one WhisperTranscribe transcription record, or all records for the authenticated user when deleteAll is set. This is destructive and can permanently remove access to the selected history entries.

delete-transcription

App-only tool used exclusively by the WhisperTranscribe widget. Models must not call it directly; call transcribe to render the widget instead. The tool prepares a bounded media upload and returns only the next-step upload payload: fileId plus upload URLs. It intentionally excludes storage keys, user identifiers, session state, telemetry, and internal diagnostics.

get-upload-url

Reads the latest job state for a user-owned WhisperTranscribe record and, when complete, returns the user-facing transcript package needed in ChatGPT: transcript text, normalized subtitle entries, summary, timestamps, language, and duration. For subtitle translation or any task that must preserve cue timestamps, use subtitleEntries when it is non-empty. Use it only when the user asks to retrieve a WhisperTranscribe transcription and a valid fileId is already known or was selected from get-filelist; when chaining from get-filelist, pass the selected record's id as fileId. Do not use it for pasted transcript/SRT text, example JSON, generic translation, or subtitle-format conversion. If the user needs the full transcription workspace or account upgrades, direct them to https://whispertranscribe.ai/transcribe or https://whispertranscribe.ai/pricing. The tool excludes personal identifiers, session state, telemetry, and internal diagnostics.

get-transcription-result

Loads the widget preference state WhisperTranscribe needs inside ChatGPT: UI language plus a coarse entitlement flag, hasPremiumAccess. It does not expose billing records, raw subscription fields, personal identifiers, session state, telemetry, or internal diagnostics.

get-user-preferences

Lists the authenticated user's recent WhisperTranscribe jobs for history, selection, and polling. Use it only for the user's WhisperTranscribe transcription history, not for files on the user's computer, Downloads folder, cloud drive, or generic file management. When a selected record is passed to get-transcription-result, map the record's id field to that tool's fileId input. The response is constrained to display-safe file fields and excludes personal identifiers, session state, telemetry, and internal diagnostics.

get-filelist

Use this tool only when the user wants WhisperTranscribe to process an audio or video source and produce a new transcription, or explicitly asks to open the WhisperTranscribe transcription interface. Do not use it to translate, edit, review, summarize, or convert transcript/subtitle text already pasted into the conversation. This tool must open the WhisperTranscribe transcription widget; do not ask the user to upload, attach, or send media files in the ChatGPT conversation because chat attachments are not processed by this app. Once the widget is visible, tell the user to choose or drag the media file inside the WhisperTranscribe widget. If the widget is not visible or cannot be opened, direct the user to https://whispertranscribe.ai/transcribe and tell them to upload the file on that page instead. After the widget is open, do not call the file-management tools such as `get-upload-url` or `insert-transcription-record` yourself: the widget uploads the media file, creates the transcription job, and polls the status automatically. Use `get-transcription-result` only when the user asks for the finished transcript data. If the user needs the full transcription workspace or account upgrades, direct them to https://whispertranscribe.ai/transcribe or https://whispertranscribe.ai/pricing.

transcribe

Updates WhisperTranscribe widget preferences stored for the authenticated user. This tool is limited to saved UI language preference and returns only the normalized preference payload, not account state, personal identifiers, session state, telemetry, or internal diagnostics.

set-user-preferences