Upload & Transcribe
Speak AI's MCP server includes tools for uploading files by URL, tracking processing, and retrieving transcripts. Use from Claude, ChatGPT, Cursor, the CLI, or any MCP-compatible client.
Upload tools
upload_and_analyze — upload a file by URL and run AI analysis in one step (recommended for most workflows)
upload_media — upload a file by URL, transcription only
upload_local_file — upload from a local file path (CLI and local MCP server only)
get_signed_upload_url — get a pre-signed URL for direct upload from your own application
Status and retrieval tools
get_media_status — check whether a file has finished processing
get_transcript — retrieve the full transcript with speaker labels and timestamps
get_captions — download the captions or subtitle file
list_supported_languages — see all 70+ supported transcription languages
Example prompts (via Claude or ChatGPT)
"Upload and transcribe this recording: [URL]"
"Check the status of media ID 12345"
"Get the transcript for my most recent upload"
"Transcribe this recording in Spanish: [URL]"
Tips
Use upload_and_analyze when you want AI insights right away — it combines upload + transcription + AI Chat in one call.
Supported formats: MP3, MP4, M4A, WAV, WEBM, MOV, OGG, and more. Max file size: 2 GB (Individual), 5 GB (Team).
Speaker identification: Virtual meetings (Zoom, Google Meet, Teams) get names automatically. Uploaded recordings get Speaker 0, 1, 2 — rename them in the app or ask your AI assistant to rename them after transcription.
Still stuck? Chat with us below or email [email protected].