Lesskeys is a native macOS app for recording, transcribing and keeping your conversations — entirely on your own machine.

Record any call, meeting, lecture or interview: system audio, your microphone and the screen or a single window, captured together as one video file. Every recording becomes an accurate transcript with automatic speaker labels, and Lesskeys learns voices so the same person is recognised across different recordings.

The part that compounds is the archive. Instead of a folder of audio files nobody opens again, you get something searchable — look inside what was actually said across every recording you own, filter by speaker, recording or date range, match an exact phrase. Click a result and the transcript opens at that line with the video in sync. Export to TXT, SRT or WebVTT.

It also does system-wide dictation: press one hotkey in any app, talk, and the text lands on your clipboard in any of 99 languages with automatic language detection. Plus a searchable clipboard manager, text replacements and per-app settings.

Nothing is uploaded. No cloud servers, no account, no subscription — the speech models sit on your disk and run on the Apple Silicon GPU, so it works with the Wi-Fi switched off. One-time purchase.

The average person uses 11 apps daily and loses 25% of their time to context switching. That's $25K wasted for every $100K of salary, moving information around instead of doing real work.

Sugarbug is a workflow intelligence platform that connects the tools you already use – Linear, GitHub, Figma, Slack, Notion, calendars, email, and more – into a single living knowledge graph. Every signal is ingested, classified, and linked automatically. Tasks, people, and the relationships between them are mapped across every source.

The longer Sugarbug runs, the smarter it gets. It builds living profiles of the people you work with from every interaction, so you always have context on who's involved in what. Meeting briefs, status updates, and cross-tool summaries are generated from real data – ready before you need them, without hunting across nine tabs.

The system is adaptive: it learns which sources matter most and adjusts how aggressively it monitors them based on actual activity patterns.

Sugarbug uses a provider-agnostic AI architecture – bring your own LLM. Pick the model that fits your needs, swap it whenever you like. No vendor lock-in.

Built for product managers, design leads, and founders who spend their days stitching together updates from half a dozen apps before they can actually do their job.