Pluely v0.1.6 is out. I'm building this in public—no VC backing, no team, just me solving problems one commit at a time.

What Pluely is: An invisible AI assistant that stays out of your way until you need it. Think Cluely, but open source, 10MB instead of 800MB, and built by someone who uses it every day.

What's different about v0.1.6: Screenshot Selection - Click and drag to capture specific areas of your screen for AI analysis. The hard part wasn't building the selection tool-it was keeping Pluely completely invisible while enabling precise interaction. I integrated tauri-nspanel on macOS for non-activating panels and rebuilt Windows window management to stay non-focusable. Three rewrites to get this right.

Audio Device Control – Pick exactly which microphone and system audio source Pluely captures from. The challenge was hot-swapping audio devices without restarting the entire capture pipeline. Had to rebuild the audio layer from scratch to support dynamic device switching.

Autostart - Pluely now launches silently when your system boots. Always ready, never in the way.

Window Focus Issues Solved - The most reported issue from v0.1.5. Pluely no longer steals focus on macOS or Windows. You stay in your workflow, Pluely stays invisible until you call it.

Custom audio capture with Voice Activity Detection

Multi-LLM support: OpenAI, Claude, Gemini, Grok, Perplexity, Mistral, Cohere, plus custom/local models

Try it: https://pluely.com

Source: https://github.com/iamsrikanthnani/pluely

Full changelog: https://github.com/iamsrikanthnani/pluely/releases/tag/app-v...

I ship updates based on what users actually need, not what looks good on a roadmap. This release exists because the community told me what was broken and what is needed.

Built in public. No hype. Just shipping.