Great Work Labs

Speaker labels, tool calls and image Q&A

Jake Marsh ยท November 19, 2026

Five new models joined the lineup, bringing it to 32.

Who splits a recording into turns and labels each speaker, with talk time per person, for $0.0003 a minute. If you want the words too, send speakers: true to Scribe and every line of the transcript comes back with its speaker, at no extra charge.

Ear listens to the first 30 seconds of a recording and names the spoken language, so you can route a voicemail or pick a transcription language before you pay for the whole file.

Align takes audio plus the script you already have and returns every one of your words with a start and end time, for word-by-word highlighting and subtitle timing.

Cue takes a request like "text mom that I'll be 10 minutes late" and your list of functions, and returns send_message with {"to": "mom", "body": "I'll be 10 minutes late"}. Arguments always match your JSON Schema, and when nothing fits you get tool: null instead of a guess. It accepts tools in OpenAI and Anthropic formats.

Glance writes alt text and descriptions for images, and answers questions about them: ask a receipt for its total and it reads it off.

RSS feed