analyze_video
ChatGPTCheck a video before dubbing: duration, whether it can be processed, and the target languages. Pass videoUrl (YouTube, X/Twitter, Instagram, Vimeo or a direct https link) or uploadId. For a file the user attached in the chat use upload_attached_file instead - never pass a /mnt/data path here. Spoken language and number of speakers are detected when the job runs.
cancel_dubbing_job
ChatGPTCancel one of the user's dubbing jobs that is still processing. Nothing is charged for a cancelled job.
create_dubbing_job
ChatGPTStart dubbing a video. This uses the user's minutes or credits: first call estimate_dubbing_cost, show the user the cost, and only call this after the user explicitly agrees. Returns immediately with the job id - tell the user straight away that the job was received; processing continues in the background, check it later with get_dubbing_status.
create_transcription_job
ChatGPTTranscribe (speech to text) a video or audio file in the background, with timestamps and optional speaker labels. Uses the user's Klassifier credits: call estimate_transcription_cost first and only call this after the user agrees. Returns immediately - tell the user the job was received, then fetch the transcript with get_transcription_result.
create_upload_link
ChatGPTLet the user upload a video or audio file from their device (shown as an upload box in the chat where supported, plus a private single-use link) (up to 4 GB) for dubbing. Use it when the video is not at a public URL. Returns an uploadId to pass to analyze_video and estimate_dubbing_cost once the user says the upload is finished.
estimate_dubbing_cost
ChatGPTEstimate what dubbing a video would cost from the user's DubbyLab minutes or Klassifier credits, and whether the balance is enough. Returns a quoteId that create_dubbing_job requires.
estimate_speech_cost
ChatGPTEstimate the length and credit cost of turning a text into speech, and whether the balance is enough.
estimate_transcription_cost
ChatGPTEstimate the credit cost of transcribing (speech to text) a video or audio file, and whether the balance is enough. Pass videoUrl or uploadId.
generate_speech
ChatGPTTurn text (up to 5000 characters) into natural speech with one of the voices from list_tts_voices. Uses the user's Klassifier credits (charged per started minute of audio) - for long texts tell the user the estimate from estimate_speech_cost first. Returns a private audio link.
get_dubbing_result
ChatGPTPrivate, time-limited download and preview links of a finished dubbing job (video or audio, subtitles).
get_dubbing_status
ChatGPTCurrent status, stage and approximate progress of one of the user's dubbing jobs.
get_supported_languages
ChatGPTList the languages videos can be dubbed into (with their languageId) and the dubbing options. The spoken source language is detected automatically.
get_transcription_result
ChatGPTStatus of a transcription job and, once completed, the transcript text (first 30,000 characters) with private download links for TXT, SRT subtitles and JSON segments.
list_my_dubbing_jobs
ChatGPTThe user's dubbing jobs, newest first.
list_my_transcriptions
ChatGPTThe user's transcription jobs started from the assistant, newest first.
list_tts_voices
ChatGPTVoices available for text to speech in a language (languageId from get_supported_languages), with gender, age, speaking style and tone. Optional filters: gender, style.
upload_attached_file
ChatGPTUse this whenever the user attaches a video or audio file in the chat (a /mnt/data file) and wants it dubbed or transcribed: pass the attached file as 'file'. It receives the file directly (no public link needed), stores it privately, and returns an uploadId plus its duration. Continue with estimate_dubbing_cost or estimate_transcription_cost using that uploadId. If the chat cannot hand the file over, the result contains an uploadUrl instead: give it to the user as a link and continue with its uploadId when they say the upload is done.