speech.transcribe

What problem does it solve?

Speech transcribe turns recorded audio or video into usable text you can search, quote, and act on, without manually listening and typing.

Core Features & Use Cases

• Provider-agnostic transcription: Uses the native audio_transcribe tool with optional provider selection (auto, OpenAI, Deepgram, or AssemblyAI) for practical reliability.
• Diarization and timestamps: Produces speaker-attributed segments (when diarization is enabled) and supports word/segment timestamps for analysis or review.
• Transcript artifacts and operational visibility: Persists transcript text and segment JSON as workspace artifacts while accounting for duration and cost, enabling controlled sharing.
• Use Case: You need a call recap with speaker-by-speaker notes and word-level timing for compliance review, generated from a single uploaded audio file.

Quick Start

Ask for the transcription by saying: transcribe the attached audio clip with diarization and word timestamps, and include the detected language.