Show HN: Nura Dev – Voice control for Claude Code, from your phone

Nura Dev is listed on the Apple App Store as a voice-control app for AI coding, marketed specifically as voice control for Claude Code from a phone. It appeared on Hacker News as a Show HN under the Claude/Codex/Fable track, with the App Store listing as the linked product page. At the time of the listing, the thread had 1 point and 0 comments, so there is no public HN discussion yet to fill in install numbers, pricing, or release history.

On the product side, the stated mechanics are narrow and concrete: the phone becomes a voice front end, and Claude Code is the coding target. That implies a remote or companion path from speech on iOS into whatever Claude Code session the developer is already running, rather than a full IDE replacement. Without published architecture notes or benchmarks, the only solid claim is the control model—hands-free, mobile-originated voice into an AI coding workflow—not latency numbers, model versions, or offline capability.

For engineers and builders, the useful angle is workflow, not novelty for its own sake. Claude Code is typically a terminal- or desktop-bound loop; a phone voice layer targets moments when the keyboard is awkward—walking a build, reviewing a long agent run, or steering a task while away from the desk. The open question for practitioners is whether the app can issue reliable, scoped commands into an active Claude Code session without constant re-auth, context loss, or noisy misrecognition that costs more time than typing.

Competitively, this sits in the thin slice of mobile companions for AI coding agents rather than general voice assistants. Claude Code already competes in the agentic coding space with tools in the same HN Claude/Codex/Fable cluster; Nura Dev is not another coding model, it is an access surface on top of one. Early Show HN traction is still essentially zero (1 point, no comments), so market signal is the product definition and App Store presence, not adoption metrics.

Practical takeaway: if you run Claude Code regularly, evaluate Nura Dev on three checks—can it attach to your real sessions, how precise is intent-to-command mapping for common ops (run, fix, explain, commit-adjacent tasks), and does phone voice stay useful once you are back at a keyboard. What to watch next is any technical write-up or demo of the phone-to-Claude Code path, plus whether the HN thread or App Store page starts showing real usage feedback beyond the initial Show HN post.