[KongchangAI]
ProductAtticus voice assistant

Atticus

Atticus is a voice assistant tool for macOS that uses a push-to-talk interaction model. It processes voice commands by invoking the Claude CLI and is designed to provide a keyboard-driven voice input experience for developers and productivity tool enthusiasts.

Timeline (last 90 days)

Oct 12

Atticus is a voice assistant tool for macOS that uses a hold-to-talk interaction method, where users press a shortcut key to start speaking and release it to send the command.

Unverified50%
Oct 12

Atticus does not build its own large model, but instead directly invokes Claude's command-line capability (claude -p), converting speech to text and then passing it to Claude for processing.

Unverified50%
Oct 12

Atticus decouples the voice input interaction layer from Claude's reasoning layer, requiring the developer only to handle voice capture, hotkey listening, and result presentation.

Unverified50%
Oct 12

Atticus depends on the user having a correctly installed and logged-in Claude CLI environment locally, making its barrier to entry skewed toward technical users

Unverified50%
Oct 12

Atticus made its debut as a project on Hacker News' "Show HN" section

Unverified50%
Oct 12

Delegating reasoning to Claude allows Atticus to benefit from Claude's continuous iteration without worrying about model hosting, compute costs, or updates, but at the cost of latency across the entire pipeline, offline availability, and limited transcription accuracy

Unverified50%

All Facts (6)

Source Articles