Product · Built by Fullspec Studio · 2026

Cue

Voice-first prompt engineering for Mac. Hold a hotkey and talk, and Cue doesn't just transcribe — it restructures your spoken prompt into a few versions to choose from before you send it to Claude, ChatGPT, or whatever you're using.

Project shape
Systems Design 8%Infra 6%DevOps 7%Data 2%AI 35%Backend 12%Frontend 30%
Leading
AI35%Frontend30%Backend12%
A MacBook Pro on a warm walnut desk in a home podcast studio, a broadcast microphone in the foreground, the screen showing a live Claude Code session with Cue's coral dictation waveform 'Listening…' along the bottom — talking a prompt straight into the terminal.

The problem

Talking is faster than typing, but raw dictation is messy — filler words, run-on sentences, half-finished thoughts. Paste that straight into an AI tool and you get a messy answer back.

What we built

Cue, a menubar app for Mac. Hold a hotkey and talk; transcription runs on-device, and then Cue doesn't just drop in the raw text — it restructures your spoken prompt using prompt-engineering techniques and hands you a few versions to choose from: cleaned, polished, a concise prompt, or a fully structured one. Pick one, tweak it if it needs it, and it drops straight into whatever field your cursor is in — Claude, ChatGPT, your editor, a Slack message.

Cue lives in the menubar — hold-to-talk dictation, Format Last Dictation, and your choice of AI provider

Transcription is on-device; the formatting step runs on Anthropic's or OpenAI's API, with your own key kept in the Mac Keychain.

We built it because we prompt AI all day, every day. Better inputs mean better outputs, and typing a careful prompt every time is slower than talking and letting the app do the first pass of structuring. Same rule as everything we build: the model drafts, you decide.

Status

Cue is in beta at cuespeech.com and launching soon. We use it daily for our own AI work — the same APIs we lean on in client automations, pointed at our own workflow first.

Work

Have a system that needs this level of attention?

See how we evaluate a new engagement.