codexstar69

codexstar69 / pi-listen

Public

Hold-to-talk voice input for Pi CLI — Deepgram streaming STT with live transcription, voice commands, and cross-platform hold detection

www.npmjs.compackage@codexstarpi-listen cli deepgram hold-to-talk pi-extension pi-package
13
2
100% credibility
Found Mar 16, 2026 at 13 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

pi-listen adds hold-to-talk voice transcription to the Pi coding agent using either cloud streaming for real-time results or downloadable offline models.

How It Works

1
🔍 Discover voice input for Pi

While using your Pi coding helper, you spot pi-listen to speak code instead of typing.

2
📦 Add the voice feature

Run a quick install command in your regular terminal to bring voice to Pi.

3
Pick your voice style
☁️
Cloud for speed

Link a free cloud service for real-time words-as-you-speak.

🏠
Offline private

Grab a model that runs fully on your computer, no internet needed.

4
🎤 Talk to code

Hold the spacebar, speak your ideas, and watch text flow into your editor like magic.

5
🛠️ Fine-tune anytime

Open the voice panel to switch languages, models, or test your mic easily.

✨ Code hands-free

Now dictate code effortlessly, speeding up your work with natural voice.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 13 to 13 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is pi-listen?

pi-listen adds hold-to-talk voice input to the Pi CLI agent, turning spoken words into typed CLI commands or code via SPACE key hold detection. Developers hold SPACE for instant transcription—live streaming with Deepgram or offline batch processing with local models—solving the hassle of typing long prompts in terminals. Built in TypeScript, it handles cross-platform audio capture with sox or ffmpeg, plus commands like /voice-settings for backend tweaks and model downloads.

Why is it gaining traction?

Dual backends stand out: Deepgram for real-time "listen to pi" streaming (56+ languages, $200 free credits) versus 19 offline models (Parakeet, Whisper) that auto-download and run on CPU. Features like pre-recording (no missed words), tail capture, continuous dictation (/voice dictate), and a unified settings panel with device-aware recommendations hook CLI users tired of keyboard-only workflows. Cross-platform hold-to-talk detection works everywhere, no fuss.

Who should use this?

CLI-heavy devs on Pi agent who dictate code or run voice commands during pair programming. Terminal hackers on macOS/Linux/Windows needing hands-free input for quick tests or notes. Power users avoiding typing in SSH sessions or low-light setups, especially with Deepgram for live hold-to-talk voice chat.

Verdict

Grab it if you're on Pi CLI—polished docs and 2-minute setup make it instantly useful despite 13 stars and 1.0% credibility score signaling early maturity. Offline mode shines for privacy, but test Deepgram first for speed; lacks broad adoption so watch for edge cases.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.