listening_convert_file
ChatGPTUse when the user says send this to Listening with an attached document. Fetches the original file and starts document-to-audio conversion. Accepts the host attachment descriptor or the file object from listening_prepare_file_upload after uploading. Never invent file identifiers, extract only a summary, or use this tool for a local path. Maximum 20 MiB. Supported formats: html, pdf, doc, docx, ppt, pptx, txt, md, rtf, epub, mobi. Supported languages: English, Spanish, French, Hindi, Italian, Japanese, Portuguese, Chinese. Check listening_get_import_options before submitting. Never submit a known unsupported language, translate to bypass restrictions, or disguise an unsupported file as text. Subscription and upload limits are enforced. Identical content and title reuse an existing library document unless its conversion failed. Returns acceptance, not ready audio. Do not automatically retry an interrupted creation; check the library first.
listening_convert_text
ChatGPTUse for pasted or selected plain text that the user asks to send to Listening or turn into audio. Submit the actual text, never a file path or an attachment summary. For attached files use listening_convert_file; for web links use listening_convert_url. Supported formats: html, pdf, doc, docx, ppt, pptx, txt, md, rtf, epub, mobi. Supported languages: English, Spanish, French, Hindi, Italian, Japanese, Portuguese, Chinese. Check listening_get_import_options before submitting. Never submit a known unsupported language, translate to bypass restrictions, or disguise an unsupported file as text. Subscription and upload limits are enforced. Identical content and title reuse an existing library document unless its conversion failed. Returns acceptance, not ready audio. Do not automatically retry an interrupted creation; check the library first.
listening_convert_url
ChatGPTUse for a user-selected public HTTPS article or document link, up to 20 MiB. For a chat attachment or prepared CLI upload use listening_convert_file. Do not submit a login page or URL requiring authorization headers. Supported formats: html, pdf, doc, docx, ppt, pptx, txt, md, rtf, epub, mobi. Supported languages: English, Spanish, French, Hindi, Italian, Japanese, Portuguese, Chinese. Check listening_get_import_options before submitting. Never submit a known unsupported language, translate to bypass restrictions, or disguise an unsupported file as text. Subscription and upload limits are enforced. Identical content and title reuse an existing library document unless its conversion failed. Returns acceptance, not ready audio. Do not automatically retry an interrupted creation; check the library first.
listening_get_document
ChatGPTCheck the conversion status of a document in the connected Listening account and return its listening link. A newly accepted conversion may still be processing. Do not repeatedly poll.
listening_get_import_options
ChatGPTCheck the connected account’s subscription, upload allowance, allowed document formats, supported source languages, and transfer limits before sending content to Listening. This does not reserve an upload; creation rechecks eligibility.
listening_list_documents
ChatGPTFind documents uploaded by the connected Listening account. Returns titles, conversion status, and listening links. Use the cursor to retrieve another page.
listening_prepare_file_upload
ChatGPTUse only when the agent has a local file and can perform HTTP uploads. Creates a temporary multipart POST destination, not an audio conversion. POST every returned field to upload_url, append the original bytes as the final field named file, then pass the returned file object to listening_convert_file. Do not paste binary data or base64 into text tools. ChatGPT attachments can go directly to listening_convert_file. Maximum 20 MiB; expires in 10 minutes. Supported formats: html, pdf, doc, docx, ppt, pptx, txt, md, rtf, epub, mobi. Supported languages: English, Spanish, French, Hindi, Italian, Japanese, Portuguese, Chinese. Check listening_get_import_options before submitting. Never submit a known unsupported language, translate to bypass restrictions, or disguise an unsupported file as text. Subscription and upload limits are enforced. Identical content and title reuse an existing library document unless its conversion failed. Returns acceptance, not ready audio. Do not automatically retry an interrupted creation; check the library first.
